PERFORMANCE EVALUATION OF WIRE ELECTRODE DISCHARGE MACHINING (WEDM) ON INCONEL 718

MOHD NIZAM BIN ALI

A project report submitted in partial fulfilment of the requirements for the award of the degree of Master of Engineering (Mechanical - Advance Manufacturing Technology)

Faculty of Mechanical Engineering Universiti Teknologi Malaysia

MAY 2010

iii

To my beloved mother and father Ali bin Mohd Jos Badariah bt. Md. oh

My beloved wife and son Roslina bt. Mamat Muhammad Ali Imran bin Mohd izam

iv

ACK OWLEDGEME T

In the ame Of Allah, the most Gracious and most Compassionate

In this opportunity, I would like to exprese my sincere appreciation to my supervisor, Professor Dr. Safian bin Sharif for the excellence supervision, guidance and continuous encouragement throughtout this study. Its also a pleasure to acknowledge to Mr. Ali and other staff of Production Laboratory, Faculty of Mechanical Engineering Universiti Teknologi Malaysia for their assistance during conducting the experiment.

I would also like to express my special thanks to my wife and son for their patience and believing in me throughout these years. Not to forget, to all friends especially Ameizza, Hassanudin, Suhaimi, Nazif and my other colleague who have contributed directly or indirectly to the success of this project.

ABSTRACT

Superalloys are known as unique materials ever produced in manufacturing industries. Its capable to withstand in high temperature and the excellent resistance in mechanical and chemical degradations. Inconel 718 is one of the superalloy material whichs is which is widely used in aeronautical and aerospace industries. This nickelbased superalloy is a high strength, thermal resistance with extreme toughness and work hardening characteristics materials. It is also noted for its excellent corrosion resistance in many conditions of engineering applications. Due to it extremely tough nature, the machinability studies of this material had been carried out by many researchers for the past few years. This master project presents the machining of Inconel 718 using wire electro-discharge machining with zinc coated brass electrode wire diameter of 0.25mm. The objective of this master project is mainly to investigate the performance of wire electro-discharge machining on Inconel 718. This is done by observing the influence of the various WEDM machining characteristics namely, surface roughness (Ra), sparking gap (Gap), material removal rate (MRR) and cutting speed (CS). A full factorial design of experiment (DOE) approach with two-level was employed to conduct this experiment. Design expert software was used to perform the ANOVA analysis and confirmation test was also conducted to verify and compare the results from the theoretical prediction using software. Overall result showed that pulse duration (ON) was the most significant factor that appeared to influence on all machining characteristics that had been investigated. The experimental results also acceptable due to the results obtain fall in acceptable values with less than 15% of margin error.

vi

ABSTRAK

Superaloi telah diketahui umum sebagai bahan yang unik yang pernah dihasilkan di dalam industri pembuatan. Ianya mampu untuk bertahan pada suhu yang tinggi dan mempunyai ketahanan yang lasak di dalam pelbagai applikasi kejuruteraan. Inconel 718 adalah salah satu bahan superaloi yang digunakan secara meluas terutamanya di dalam industri aeronatikal dan angkasa lepas. Superaloi berasaskan bahan nickel ini mempunyai kekuatan, rintangan haba dan ketahanan karat yang tinggi serta dicirikan juga dengan pengerasan kerja yang baik. Berdasarkan kepada sifat Inconel 718 yang tahan lasak, kajian kebolehmesinan bahan ini telah menjadi minat pengkaji sejak beberapa tahun kebelakangan ini. Projek sarjana ini bertujuan untuk menyiasat prestasi kebolehmesinan Inconel 718 menggunakan proses pemotongan nyahcas-elekrik menggunakan wayar elektrod tembaga bersalut zink berdiameter 0.25mm. Ianya melibatkan ujikaji serta pemerhatian terhadap ciri-ciri pemesinan Inconel 718 seperti kekasaran permukaan (Ra), jarak percikan api (Gap), kadar pemotongan bahan (MRR) dan kelajuan pemotongan (CS). Rekabentuk ujikaji dengan pendekatan full factorial dua tahap telah digunakan di dalam ujikaji ini. Perisian Design Expert juga telah digunakan untuk tujuan analisa varian (ANOVA) bagi setiap keputusan ujikaji. Bagi tujuan penentuan ralat, ujikaji pengesahan dilaksanakan untuk menguji kesahihan dan perbandingan diantara keputusan yang dihasilkan oleh ujikaji dan juga secara teori. Secara keseluruhannya, keputusan ujikaji menunjukkan tempoh masa denyutan (ON) adalah faktor yang paling signifikan mempengaruhi semua ciri pemesinan yang dikaji. Data ujikaji juga menunjukkan perbezaan margin di bawah nilai 15% dan ianya adalah didalam julat yang boleh diterima pakai di dalam analisa ini.

vii

CO TE TS

CHAPTER

TITLE

PAGE

DECLARATIO DEDICATIO ACK OWLEDGEME TS ABSTRACT ABSTRAK CO TE TS LIST OF TABLES LIST OF FIGURES LIST OF APPE DICES

ii iii iv v vi vii xi viii xv

O E

I TRODUCTIO

1.1 1.2 1.3 1.4 1.5

Project Background & Rationale Research Statement Research Objectives Scope of Study Expected Results

1 3 4 4 5

viii TWO LITERATURE REVIEW

2.1 2.2 2.3

Introduction Wire Electrical Discharge Machining (WEDM) WEDM Machining Parameter 2.3.1 2.3.2 2.3.3 2.3.4 Pulse duration (On-time) Pulse interval (Off-Time) Servo voltage Peak current

6 7 12 13 14 14 14 15 16 17 19 20 21 21 22 24

2.4

Machining Characteristic 2.4.1 2.4.2 2.4.3 Effect on surface finish, (Ra) Effect on material removal rate, MRR Effect on spark gap, Gap

2.5

Wire Electrode 2.5.1 2.5.2 2.5.3 Copper wire Brass wire Coated wire

2.6

Nickel Based Superalloy and Their Machinibility. 2.6.1 Inconel 718 physical properties and mechanical properties

26 27 28

2.7

Design of Experiment (DOE) 2.7.1 Two-level full factorial design

THREE

METHODOLOGY

3.1 3.2 3.3

Introduction Research Design and Data Analysis Research Design Variable 3.3.1 3.3.2 Machining parameters Machining characteristic

30 31 33 33 36

ix 3.3.3 3.3.4 3.3.5 3.3.6 3.4 3.5 3.6 Surface roughness, Ra Sparking gap, Gap Cutting speed, Cs Material removal rate, MRR 36 37 37 37 38 41 43

Research Procedure Experimental Set-up Experimental Equipment

FOUR

EXPERIME TAL RESULTS A D DATA A ALYSIS

4.1 4.2 4.3

Introduction Experimental Results Analysis of Results 4.3.1 4.3.2 4.3.3 4.3.4 Analysis results for surface roughness, Ra Analysis results for sparking gap, Gap Analysis results for material removal rate, MRR Analysis results for cutting speed, CS

46 47 54 54 58 63 68 74 75 77 79

4.4

Confirmation Test 4.4.1 Confirmation test and results

4.5 4.6

Comparisons of the Test Results Verification of Mathematical Models

FIVE

DISCUSSIO S

5.1 5.2 5.3 5.4 5.5

Introduction Surface Roughness, Ra Sparking Gap, Ra Material Removal Rate, MRR Cutting Speed, CS

83 84 85 86 88

x SIX CO CLUSIO S A D RECOMME DATIO S

6.1 6.2

Conclusions Recommendations

90 91

REFERE CE APPE DIX A APPE DIX B APPE DIX C APPE DIX D

92 99 102 105 108

xi

LIST OF TABLES

O.

TITLE

PAGE

2.1 2.2 2.3 2.4 3.1 3.2 3.3 3.4 4.1 4.2 4.3 4.5 4.6 4.7 4.8 4.9 4.10 4.11 4.12 4.13

The designation of various types of coated wire and their applications Chemical properties of Inconel 718 (%) Physical properties of Inconel 718 Mechanical properties of Inconel 718 Two-level full factorial experiment with four factors The design of machining parameters Actual value of experiment design Machining parameters set-up (constant parameters) Experimental results of surface roughness (Ra) Experimental results of sparking gap (Gap) Experimental results for cutting speed (CS) Experimental results for material removal rate (MRR) Overall experimental results corresponded to each run ANOVA for surface roughness (Ra) ANOVA for sparking gap (Gap) ANOVA for material removal rate (MMR) ANOVA for cutting speed (CS) Summary of the significant factors in WEDM Inconel 718 Quality characteristic of the machining performance True value of confirmation test experiment

23 26 27 27 32 34 35 41 48 49 50 52 53 55 59 63 68 74 75 75

xii O. TITLE PAGE

4.14 4.15 4.16 4.17 4.18 4.19 4.20 4.21 4.22

Confirmation test results for surface roughness (Ra) Confirmation test results for sparking gap Confirmation test results for cutting speed (CS) Confirmation test results for material removal rates (MRR) Comparison test results for surface roughness (Ra) Comparison test results for sparking gap (Gap) Comparison test results for material removal rate (MRR) Comparison test results for cutting speed (CS) Margin of error for actual results and predicted values (%)

76 76 76 76 78 78 78 78 82

xiii

LIST OF FIGURES

O.

TITLE

PAGE

2.1

Typical product by WEDM (AGIE Charmilles groups, Charmilles the solution when to EDM, Geneva 2004) 7 8 9

2.2 2.3 2.4

Classification of EDM processes Wire electrical discharge machining (WEDM) processes Schematic of the thermal removal processes of WEDM (Spark between electrode and workpiece perform the material removal)

10 11 19 31 40 44 44 45 56 57 57 60 61 61 65

2.5 2.6 3.1 3.2 3.3 3.4 3.5 4.1 4.2 4.3 4.4 4.5 4.6 4.7

Classification of major EDM research areas Definition of kerf and overcut in WEDM Research methodology Flowchart of experiment steping WEDM linear motor 5 axes Sodick series AQ537L Mitutoyo surface roughness measuring machine Zeiss Axiotech high power optical microscope Normal probability plot of residuals for surface roughness (Ra) Residual vs. predict response for surface roughness (Ra) Main effect plot for surface roughness (Ra) Normal probability plot of residuals for sparking gap (Gap) Residual vs. predict response for sparking gap (Gap) Main effect plot for sparking gap (Gap) Normal probability plot of residuals for material removal rate (MRR)

xiv O. TITLE PAGE

4.8 4.9 4.10 4.11 4.12 4.13 4.14 5.1 5.2 5.3 5.4 5.5

Residual vs. predict response for material removal rate (MRR) Main effect plot for material removal rate (MRR) Interaction between SV*OFF for material removal rate (MRR) Normal probability plot of residuals for cutting speed (CS) Normal probability plot of residuals for cutting speed (CS) Main effect plot for cutting speed (CS) Interaction plot of cutting speed (CS) 3D interaction graph for surface roughness (Ra) 3D interaction graph for sparking gap (Gap) 3D interaction graph for material removal rate 3D interaction graph of IP*SV for cutting speed (CS) 3D interaction graph of OFF*SV for cutting speed (CS)

65 66 67 70 70 71 72 84 85 87 88 89

xv

LIST OF APPE DICES

O.

TITLE

PAGE

A1 A2 B1 C1 C2 D1 D2

Schedule for Master project part I (Semester 1 2009/2010) Schedule for Master project part I (Semester 1 2009/2010) Summary of finding related to EDM performance Experimental results of sparking gap (top surface) Experimental results of sparking gap (bottom surface) Confirmation experimental results of sparking gap (top surface) Confirmation experimental results of sparking gap (bottom surface)

100 101 103 106 107 109 109

CHAPTER O E

I TRODUCTIO

1.1

Project Background and Rationale

Electrical discharge machining (EDM) is a non-traditional concept of machining which has been widely used to produce dies and molds. This technique has been developed in the late 1940s and has been one of the fast growing methods in manufacturing during 1980s and 1990s [1].

This non-traditional machining method is commonly used for very hard metals that would be impossible to machine with traditional techniques. It has been extensively used, especially for cutting intricate contours or delicate cavities that also would be difficult to produce with a conventional machining methods or tools. However, one critical limitation is that EDM only works with electrically conductive materials. Metal that can be machined by using EDM include nickel-based alloy (such as inconel), hardened tool steels and carbides.

Wire electrical discharge machining (WEDM) is introduced in the late 60s. The process was fairly simple, not complicated and wire choices were limited to copper and brass only. WEDM is a thermo-electrical process in which material is

eroded from the workpiece by a series of discrete sparks between the workpiece and the wire electrode (tool) separated by a thin film of dielectric fluid (deionized water) that is continuously fed to the machining zone to flush away the eroded particles. The movement of wire is controlled numerically to achieve the desired three-dimensional shape and accuracy of the workpiece [2]. The degree of accuracy of workpiece dimensions obtainable and the fine surface finishes make WEDM particularly valuable for applications involving manufacture of stamping dies, extrusion dies and prototype parts. Without WEDM, the fabrication of precision workpieces requires many hours of manual grinding and polishing [3].

In recent years, the technology of wire electrical discharge machining (WEDM) has been improved significantly to meet the requirements in various manufacturing needs, especially in the precision mold and die industry. WEDM is being used to machine a wide variety of miniature and micro-parts from metals, alloys, sintered materials, cemented carbides, ceramics and silicon. This tremendous achievement in WEDM technology has been achieved by many researchers from some of the world leading institution and research centre, but still cannot coped with the new materials introduced to the market.

The selection of cutting parameters for obtaining higher cutting efficiency or accuracy in WEDM is still not fully solved, even with the most up-to-date CNC WEDM machine. This is mainly due to the nature of the complicated stochastic process mechanisms in wire-EDM. As a result, the relationships between the cutting parameters and the process performance are hard to achieve accurately [4]. There is still lack of research on WEDming of material such as nickel based super alloy which include Inconel 718. It is widely used; mostly in aerospace and marine applications which are classified as difficult to machine material by conventional method due to high cutting temperature and rapid tool wear [5].

1.2

Research Statement

Studies on WEDM using coated wire somehow is limited and manufactures claimed that the outstanding performance is achieved through their lab test. But the results are not disclosed to the public and researcher for further study and understanding. As such the machine materials information and the WEDM parameters setting for the subjected wire are somehow limited. The only information given/set by manufactures is commonly applicable to the common steel grades [6].

Inconel 718 is a high strength and thermal resistance (HSTR) [7] known to play increasingly important in the aviation, space navigation and shipping industries because of its outstanding multi-properties [8]. Broad bases of Inconel 718 knowledge are now exist due to its great acceptance in industries. However, the parameter setting on WEDM of Inconel 718 is still lacking. The available technological data which is based on manufacturers for in house experimentation is helpful but insufficient.

Inconel 718 is assigned to be machined with WEDM in this project with the attention to study the parameters setting for an optimum machining. A comparative study will be carried out between previous study using brass wire and the proposed study using coated wire electrode.

1.3

Research Objectives

The objectives of the research are:

a.

To determine the significant parameters that influences the machining responses during Wire Electro-Discharge Machining (WEDM) of Inconel 718.

b.

To evaluate the performance of Electro-Discharge Machining (WEDM) on Inconel 718 with respect to various responses such as spark gap, material removal rate, cutting speed and surface finish.

c.

To establish mathematical model for spark gap, surface finish, cutting speed and material removal rate during WEDM of Inconel 718 alloy.

1.4

Scope of Study

The scope of the research consists of:

a.

Wire Electro-Discharge Machining, (WEDM) linear motor 5-axis Sodick series AQ537L will be employed.

b. c. d.

Nickel based superalloy, Inconel 718 will be used as the workpiece material. Zinc coated brass wire of diameter 0.25mm will be used as electrode. Parameters to be studied include voltage, peak current, pulse duration and interval time.

e.

Response variable to be study are surface finish, spark gap, material removal rate and cutting speed.

f.

The DOE and analysis of variance (ANOVA) will be processed using Design Expert software version 7.0.0.

1.5

Expected Results

The expected outcomes of this study are as follows:

a.

To obtain the optimum condition for WEDM on Inconel 718 in various parameters setting using zinc coated brass wire.

b.

Establishment of mathematical models for various responses during WEDM on Inconel 718

c.

The outcome of the study can be used to assist the industrial practitioners that involved in machining of superalloy materials such as nickel alloys and to select the most suitable cutting parameters for machining nickel alloys application.

d.

This will help in improving the quality of Inconel products as well as minimizing the machining cost to realize the economical potential to the fullest.

CHAPTER TWO

LITERATURE REVIEW

2.1

Introduction

Wire electrical discharge machining (WEDM) is a metalworking process with the help of which a material is separated from a conductive work piece, by means of rapid, repetitive spark discharges from a pulsating direct-current power supply with dielectric flow between the workpiece and the tool [8]. Research in areas WEDM has become a considerable interest due to the various advantageous offered by this process. Among the various non-conventional machining processes, WEDM is the most widely and successfully used method for machining difficult to machined materials such as super alloys [9].

Considering the increasing number of high-strength, non-corrosion and wear resistant materials such as Inconel 718, WEDM has brought many improvements in recent years. Researchers are struggling to reveal a new method to improve WEDM efficiency; the objectives are the same: to enhance the capability of machining performance, to get better output product, to develop technique to machine new materials and to have better working conditions [1]. This is due to WEDM

technologies offers no readily available standard for setting the cutting parameters

7 such as current, polarity, duty cycle, etc, [9] to achieve the desire machining characteristics of the nickel alloys in particular Inconel 718.

The selection of cutting parameters for obtaining higher cutting efficiency or accuracy in WEDM is still not fully solved, even with the most up-to-date WEDM machine. This is mainly due to the nature of the complex stochastic process mechanisms in WEDM [10].

2.2

Wire Electrical Discharge Machining (WEDM)

Wire electrical discharge machining (WEDM) has been found to be an extremely potential electro-thermal process in the field of conductive material machining. Owing to high process capability it is widely used in manufacturing of cam wheels, special gears, bearing cage, various press tools, dies, and similar intricate parts etc (Figure 2.1) [11].

Figure 2.1: Typical product by WEDM (AGIE Charmilles groups, Charmilles the solutions when to EDM, Geneva, 2004.)

8 According to previous researcher, Sommer [12] EDM can be categorized into two: die sink EDM and wire EDM. Pandey and Shah [13] classified EDM processes into three main categories as shown in Figure 2.2. EDM techniques have developed in many areas. Trends on activities carried out by researchers depend on the interest of the researchers and the availability of the technology. In 1994, Rajurkar [14] has indicated some future trends activities in EDM: machining advanced materials, mirror surface finish using powder additives, ultrasonic-assisted EDM control and automation.

Figure 2.2: Classification of EDM processes [13]

The concept of WEDM is shown in Figure 2.3. In this process, a slowly moving wire travels along a prescribed path and removes material from the workpiece. WEDM uses electro-thermal mechanisms to cut electrically conductive materials. The material is removed by a series of discrete discharges between the wire electrode and the workpiece in the presence of die-electirc fluid, which creates a path for each discharge as the fluid becomes ionized in the gap. The area where discharge takes place is heated to extremely high temperature, so that the surface is melted and removed. The removed particles are flushed away by the flowing dielectric fluids as shown in Figure 2.4. The taper can ranging from 15 for a 100mm thick to 30 for a 400mm thick workpiece can be obtained on the cut surface material [15].

The wires for WEDM are made of brass, copper, tungsten, molybdenum (0.05 0.3mm in diameter) which capable to achieve very small corner radii. Zinc or brass coated wires are also used extensively in this process. The wire used in WEDM process should posses high tensile strength and good electrical conductivity.

Figure 2.3: Wire electrical discharge machining (WEDM) process [16].

10

Figure 2.4: Schematic of the thermal removal process of WEDM (spark between the electrode and workpiece perform the material removal) [17].

WEDM process is usually used in conjunction with CNC and will only work when a part is to be cut completely through. The melting temperature of the parts to be machined is an important parameter for this process rather than strength or hardness. The surface quality and MRR of the machined surface by wire EDM will depend on different machining parameters such as applied peak current, and wire materials. WEDM process is commonly conducted on submerged condition in a tank fully filled with dielectric fluid; nevertheless it also can be conducted in dry condition. This method is used due to temperature stabilization and efficient flushing in cases where the workpiece has varying thickness [18].

Although both conditions (submerged or dry machining) can be performed, most important is to produce a good quality of machined surface and dimensional accuracy. The main goals of WEDM manufacturer and users are to achieve a better stability and higher productivity of the WEDM process. As newer and more exotic materials are developed, and more complex shapes are required, conventional machining operation will continue to reach their limitations and the increased use of WEDM in manufacturing will continue to grow at an accelerated rated [19]. However, due to a large number of variables in WEDM, it is difficult to achieve the optimal performance of WEDM processes [20] and the effective way of solving this

11 problem is to establish the relationship between the performance measures of the process and its controllable input parameters.

Ho and Newman [6] have classified research areas in EDM machining process as shown in the Figure 2.5. Investigation into the influences of machining input parameters on the performance of WEDM have been widely reported [1, 4, 8, 10]. Several attempts have been made to develop mathematical model of the process [4, 11]

Figure 2.5: Classification of major EDM research areas [6]

In this project, focusing is more on improving the performance measures including MRR, spark gap and surface finish. These responses are mainly depend on the discharge energy, electrical pulse parameters and discharge distribution in space and time and flushing condition [21]. Scot et al. [22] found that current, pulse duration and pulse frequency were the main significant control factors for both the MRR and surface finish. Whereby, wire speed, wire tension, and dielectric flow were relatively significant. In addition, Ahmet Hascalyk and Ulas Caydas [23] reported that, surface roughness primarily depend on pulse duration, open circuit voltages and dielectric fluid pressure and wire speed not seeming to have much influence.

However, WEDM also have some limitation associated during machining processes including the use of electrically conductive material only, also, material removal rates are very low as compared to other processes and the work surface layer

12 is damaged after processing with this technique [6]. In addition, the selection of the appropriate parameters is also difficult and relies heavily on the operators experienced and machining parameters tables provided by the WEDM machine builder [24].

2.3

WEDM Machining Parameters

According to Wang and Yan [25], EDM parameters consist of two functional group:

a.

Electrical Parameters: polarity, peak current, pulse duration, power supply voltage.

b.

Non-electrical Parameters: rotational of speed electrode, injection flushing pressure.

Van Tri [26] categorized the EDM parameters into five groups:

a. b.

Dielectric fluid - type of dielectric, temperature, pressure, flushing system. Machine characteristics - servo system and stability stiffness, thermal stability and accuracy.

c. d. e.

Tool - material, shape, accuracy Workpiece. Adjustable parameters - discharge current, gap voltage, pulse duration, polarity, charge frequency, capacitance and tool materials.

Other studies have been carried out in order to determine the significant WEDM machining parameters that affect the performance of the WEDM processes. According to Mas Ayu [27], WEDM machining parameters had more significant effect than the electrical parameters. Her finding concluded that the most significant WEDM machining parameters are pulse duration, voltage, peak current and interval

13 time. Liao et al. [24] proposed the significant factors affecting the machining performance are spark frequency, average gap voltage and ratio of normal sparks to total sparks. Nihat Tosun et al. [28] described the highly effective parameters on both kerf and MRR were found as open circuit voltage and pulse duration whereas wire speed and dielectric flushing pressure were less effective factors. Previous researchers findings indicate that the electrical parameters are more significant than non-electrical parameters on the machining characteristic.

Several attempts also have been carried out by many researchers to investigate the effect of non-electrical parameters on WEDM machining characteristic. Erden [29] reported that dielectric flushing affected the EDM performance due to the changing of erosion rate, mirror like finishing achieved by multi divided electrode method. Kinoshita et al. [30] proved dielectric pressure greatly influence the WEDM parameters during recuts.

Furthermore, trend on activities carried out by researchers depends on the interest of the researchers and the availability of the technology. Rajurkar [14] has indicated some future trends activities in EDM: machining advanced materials, mirror surface finish using powder additive, ultrasonic-assisted EDM, control and automation.

2.3.1 Pulse duration (On-Time)

During WEDM all the work is done during pulse duration (On time). The erosion rates are affected mainly by pulse parameter. The spark gap is bridged, current is generated and the work is accomplished. The longer the spark is sustained, the higher is the material removal. Consequently the resulting craters will be broader and deeper; therefore the surface finish will be rougher. Obviously with shorter duration of sparks the surface finish will be better.

14 2.3.2 Pulse interval (Off-Time)

While most of the machining takes place during on time of the pulse, the off time during which the pulse rests and the re-ionization of the die-electric takes place, can affect the speed of the operation in a large way. Longer is the off time greater will be the machining time. But this is an integral part of the EDM process and must exist. The off time also governs the stability of the process. An insufficient off time can lead to erratic cycling and retraction of the advancing servo, slowing down the operation cycle. In addition, the interval time also provides the time to clear the disintegrated particles from the gap between the electrode and workpiece for efficient cut removal [1]. Too short pulse interval will increase the relative wear ratio and will increase the surface roughness of the machine surface [31].

2.3.3 Servo voltage

The preset voltage determines the width of the spark gap between the leading edge of the electrode and the work piece. High voltage settings increase the gap and hence the flushing and machining. Some material may be necessary for high openopen voltage due to high electrical resistance and high discharge voltage [1, 27].

2.3.4 Peak current

Peak current is also another important primary input of WEDM process. The stronger the discharge current, MRR, overcut and surface roughness will increase. In other hand, decrease the rate of electrode wear [31]. To minimize the electrode wear

15 and keep the current density between the tolerance limit it is necessary to select an appropriate value of current [18, 28].

2.4

Machining Characteristic

WEDM performance is mainly measured by the material removal rate (MRR), spark gap (kerf) and surface roughness of the workpiece. This three machining characteristics have been identified by the previous researchers as the most significant machining criteria that can influence the WEDM performance [22 - 24]. Determining the optimum machining parameters of WEDM to machine certain material is very crucial. Thicker oxide layer formed due to thermal oxidation during WEDM process is expected [31] and can reflect the surface finish of the machined surface.

Any machined surface during machining processes will experiences a disturbed layer which has different characteristics from those of the base metal. Surface integrity entails the study and control of surface topography, as well as surface metallurgy [9]. O.A. Abu Zeid [32] claimed, that thermal nature of the WEDM process always produce a recast and underlying heat-effected zone on the surface being machined and develops a residual stress that often causes micro cracks. Thermal sensitivity or chemical complexity of the material can also affect the surface integrity [23]. Several other studies also have been carried out to determine the appropriate EDM machining parameters combination from the aspect of surface integrity. The surface crack formation for AISI D2 and H13 has been studied by H.T. Lee & T.Y. Tai [33]. It was reported that crack formation and white layer thickness are related to the EDM parameters.

Ahmet et. al, [23] concluded surface integrity can be divided into two important categories; surface texture, which concern principally on the surface roughness and surface metallurgy, which concern to the nature of the surface layer

16 produced during machining. Hence, the selection of the machining parameters including pulse-on time, pulse off-time, table feed rate, flushing pressure, wire tension, wire velocity, etc should be chosen properly according to the workpiece properties so that better performance can be obtained [23]. However based on the previous findings, the significant selection of the appropriate machining responses for cutting Inconel 718 using WEDM for this project is surface roughness (Ra), spark gap (Gap) and material removal rate (MRR).

2.4.1 Effect on surface finish, Ra

Surface topography or surface finish, also known as surface texture are terms used to describe the general quality of machined surface, which is concerned with the geometric irregularities and the quality of a surface [9]. The quality of a machined surface is becoming more and more important to satisfy the increasing demands of sophisticated component performance, longevity, and reliability.

Fine surface finish is obtained by a combination of the proper electrode material, good flushing conditions, and the proper power supply settings. High frequency, low power and orbiting produce the best finish, as these conditions produce smaller, less defined craters in the work metal [8, 11, 27, 28]. Pandey and Shah [13] have found that surface finish to be inversely proportional to the frequency of discharge. Assuming that each spark leads to a spherical crater formation on the surface of workpiece, the volume of metal removed per crater will be proportional to the cube of the crater depth. Zhang et al. [34] has investigated the effect of discharge voltage, discharge current and pulse duration on AISI 1045 steel as workpiece. The investigation revealed that surface finish increase with an increase of this factor.

In 2005, M. S. Hewidy et al. [4] has investigated the WEDM performance on Inconel 601. This work has been established based on the response surface methodology (RSM). They have confirmed surface roughness increase with the

17 increase of peak current and decrease with the increase of duty factor and wire tension. Many researchers concluded that the ideal surface finishes are rare to happen due various factors. Zhang et al. [34] developed a theoretical model to estimate the surface roughness. Investigation have been carried out using AISI 1045 steel as work piece material and copper as the electrode. Results showed that the roughness of finished surface increases with an increase in the discharge voltage, discharge current and pulse duration. Guo et al. [39] also concluded that with ultrasonic aid the cutting efficiency of WEDM can improve the surface finish quality.

2.4.2 Effect on material removal rate, MRR

The removal of material in electrical discharge machining is based upon the erosion effect of electric sparks occurring between two electrodes. Several theories have been forwarded in attempts to explain the complex phenomenon of "erosive spark". The following are the theories:

a. b. c.

Electro-mechanical theory Thermo-mechanical theory Thermo-electric theory

Electro-mechanical theory suggests that abrasion of material particles takes place as a result of the concentrated electric field. The theory proposes that the electric field separates the material particles of the workpiece as it exceeds the forces of cohesion in the lattice of the material. This theory neglects any thermal effects. Experimental evidence lacks supports for this theory.

Whereby, thermo-mechanical theory suggests that material removal in EDM operations is attributed to the melting of material caused by "flame jets". These so called flame jets are formed as a result of various electrical effects of the discharge.

18 However, this theory does not agree with experimental data and fails to give a reasonable explanation of the effect of spark erosion.

Thermo-electric theory is best-supported by experimental evidence, suggests that metal removal in EDM operations takes place as a result of the generation of extremely high temperature generated by the high intensity of the discharge current. Although well supported, this theory cannot be considered as definite and complete because of difficulties in interpretation.

Material removal rate is proved by the previous researchers as one of the most important output parameters, which decide the performance of WEDM machining processes [1, 2, 4, 9, 11]. Rival [9] have discovered factors such as current, voltage, pulse on time and interval time which to have significant effect on MRR and EWR. Most researchers also concluded EDM electrical parameters such as polarity, peak current, pulse duration and power supply voltage are highly influence the MRR for performance of EDM processes [20, 21, 24, 27].

In 1991, Kunieda et al. [40] has revealed a new method to improve EDM efficiency by supplying oxygen gas into this gap. This new method results shows, material removal rate is increased due to the enlarged volume of discharged crater and more frequent occurrence of discharge. Powder additive method also has been carried out by previous researcher in order to improve the EDM efficiency. Jeswani [41] revealed that the addition of about 4 g/l of fine graphite powder in kerosene increases MRR by 60% and tool wear by 15%.

19 2.4.3 Effect on spark gap, Gap

The workpiece and wire electrode represent positive and negative terminal DC electric circuit and separated by a controlled gap which constantly controlled by the machine [31]. This gap is fulfill with dielectric fluid which act as insulator, cooling as well as flushing agent in order to flushed away the eroded particles from the cutting zone (Figure 2.6).

Figure 2.6: Definition of kerf and overcut in WEDM [42]

Spark gap is the most crucial parts of the EDM system. The size of the gap is governed by the servo control system whose motion is controlled by gap width sensors. They control the motion of the ram head or the quill which in turn governs the gap size. Spark gaps in WEDM make the kerf larger than the wire diameter as shown in Figure 2.6. This overcut is in the range of 0.020 0.050 mm [42]. The most common sparking gap is 0.03 mm. Once cutting condition has been establish for a given cut, the overcut remains fairly constant and predictable.

In this project, water is used as the dielectric. During WEDM processes, sparking occurs between the side and machine surface of the workpiece. The sparking area consists only the front of the electrode diameter (180) as it progress into the cut while the clearance is equal to the spark length of the wire electrode [31]. Sparks are formed through a sequence of rapid electrical pulse, generated by the

20 WEDM machine power supply thousands of times per second. Each spark forms an ionization channel under extremely high heat and pressure resulting in vaporization of localized sections. The vaporized metallic debris created by this process, from both the workpiece and wire material, is subsequently quenched and flushed away by the flow of dielectric fluid through the gap [43].

Literature studies showed there is fewer researchers investigated the correlation between machining parameters and spark gap in the WEDM process. Nihat Tosun et al. [28] investigated the correlation between the machining parameters and spark gap as a factor in determining the WEDM performance. The results concluded that open circuit voltage and pulse duration have the significant impact to both MRR and kerf width. Whereby, wire speed and dielectric pressure were less effective factors. Appendix B shows the summary of the researches that have been done in evaluating the WEDM machine performance.

2.5

Wire Electrode

Previous researchers M.S. Hewidy et al. [4]; S. Sarkar et al. [36] and R. Ramakrisnan et al. [38] used brass wire as the electrode to WEDM the workpiece. Brass wire is widely used in WEDM processes due to its good machining properties and can be die casted or extruded for specialized application. It possesses high tensile strength, high electrical conductivity, and good wire drawability to close tolerances.

Other researchers such as S.S Mahapatra and Amar Parnaik [35]; Kuang Yuan and Ko Ta Chiang [37] used coated wire electrode to investigate WEDM machining performance. Coated brass wire can perform at higher cutting speed compared to brass wire electrode. Coated brass wire also can produce exceptional surface finish especially when WEDM tungsten carbide and often utilized for cutting PCD and graphite.

21 The ideal wire electrode material for this process has three important criteria: a. b. c. High electrical conductivity. Sufficient mechanical strength. Optimum sparks and flushes characteristics.

As discussed above, there is no perfect wire that excels in every criteria, and some compromises become necessary, depending upon the desired results and application. And all the three factors are very closely related and interdependent.

2.5.1 Copper wire

Copper was the original material first used in WEDM. It is an excellent conductor with 100 IACS (International Annealed Copper Standard) value [31].
Although its conductivity rating is excellent, its low tensile strength, high melting point and low vapor pressure rating severely limited its potential. Under the electro-thermal condition, predominant during WEDM, copper wire wears rapidly and its tension ability is rather poor, resulting, therefore, in machining instabilities, due to high degree of short circuits, especially in the machining of small curvature [27].

2.5.2 Brass wire

Brass was the first logical alternative to copper when early EDM researchers were looking for better performance. Brass EDM wire is a combination of copper and zinc, typically alloyed in the range of 6365% Cu and 3537% Zn [43].

22 The addition of zinc element provides significantly higher tensile strength, lower melting point and higher vapor pressure rating, which more than offsets the relative losses in conductivity. Brass quickly became the most widely used electrode material for general purpose wire EDM. It is now commercially available in a wide range of tensile strengths and hardness.

2.5.3 Coated wire

Coated wire is commonly employed in WEDM process to increase substantially

the cutting speed and cutting precision. Since brass wires cannot be efficiently fabricated with any higher concentration of zinc, the logical next step was the development of coated wires, sometimes called plated or stratified wire. They typically have a core of brass or copper, for conductivity and tensile strength, and are electroplated with a coating of pure or diffused zinc for enhanced spark formation and flush characteristics.

Originally called speed wire due to their ability to cut at significantly higher metal removal rates [2], coated wires are now available in a wide variety of core materials, coating materials, coating depths and tensile strengths, to suit various applications and machine requirements. Although more expensive than brass, coated wires currently represent the optimum choice for top all-around performance, and their relative economics are covered in a later section. Table 2.1 indicates the designation of these coated wire and their applications [31].

23 Table 2.1: The designation of various types of coated wire and their applications Types, used / Charmilles Designation 1. Zinc coated brass: Half hard zinc coated brass ( 0.20 0.25mm), SS20-SS25. 2. Diffused zinc coated copper ( 0.25 0.30mm), XS25-XS30 3. Brass coated with special alloy ( 0.30 0.07mm) Special alloy Cu CuZn37

Basic Material

Applications

Characteristic and application similar to those of half hard brass wire Better usage for cylindrical cut, 1 roughness and 2 finishing passes Allows high wire tensions making it possible to produce high precision punches and dies with fines detail.

Based on the literature and considering all the important criteria, zinc coated brass wire will be employed to machine Inconel 718 in this investigation. Zinc coated brass wire was one of the first attempts to present more zinc to the wires cutting surface. This wire consists of a thin (approximately 5 m) zinc coating over a core which is one of the standard EDM brass alloys [43]. This wire offers a significant increase in cutting speed over plain brass wires, without any sacrifice in any of the other critical properties. Zinc coated brass wires produce exceptional surface finishes when cutting tungsten carbide and are often utilized for cutting PCD and graphite. These wires are also utilized in those circumstances in which brass wires produce unacceptable brass plating on the workpiece.

Due to the low electrical conductivity of Inconel 718, zinc coated brass wire is suitable choice due to its high electrical conductivity. IACS (International Annealed Copper Standard) number is one of the methods to identify the types of wire in accordance to its conductivity percentage. Copper has known to be excellent conductor with 100 IACS value. Brass alloy 63%Cu + 37% Zn = (brass) has 29

24 IACS value, while molybdenum wire has 34 IACS value. As for these projects zinc coated with copper or brass wire was chosen as it has 84 IACS value.

Zinc coated high conductivity copper alloy offers a number of superiority; high temperature toughness, high current efficiency and high discharge performance make it the best electrode wire for high speed, high precision fine machining. The core is made of copper-alloy, therefore good in workability and superior straightness which is optimum for automatic wire connection.

2.6

ickel Based Super Alloy and Their Machinibility

The number of superalloys that have been developed and used over the years is large. In reality, the solid solution strengthened alloys are strengthened both by solid solution hardening and by the presence of carbides, while the precipitation hardened alloys are strengthened by the combination of precipitates, solid solution hardening and the presence of carbides.

Nickel based superalloys are the most complex of the superalloys and are used in the hottest parts of aircraft engines, constituting over 50% of the engine weight. They are either solid solution hardened for lower temperature use or precipitation hardened for higher temperature use. The nickel based alloys contain at least 50% nickel and are characterized by the high phase stability of the FCC austenitic () matrix. Many nickel based alloys contain 1020% chromium, up to about 8% aluminum and titanium combined 515% cobalt, and small amounts of boron, zirconium, hafnium and carbon. Other common alloying additions are molybdenum, niobium, tantalum, tungsten and rhenium. Chromium and aluminum are important in providing oxidation resistance by forming the oxides Cr2O3 and Al2O3 respectively [46].

25 The most commercially superalloy is Inconel 718, listed as an ironnickel based alloy even though it contains more nickel than iron. This classification fits with the traditional classification for this alloy, although many newer works list it as a nickel based alloy. Also note that for the cobalt based alloys, they are none listed as being precipitation hardened, because unfortunately, these alloys do not precipitation harden like the nickel and ironnickel alloys.

Nickel based superalloy namely; Inconel 718 is difficult to machine material [7], perhaps second only to titanium in machining difficulty, although those who machine Inconel 718 would probably maintain that these superalloys are the most difficult to machine material. Many of the same characteristics that make Inconel 718 good high temperature materials also make them difficult to machine, namely [46,47]:

a. b. c.

Retention of high strength levels at elevated temperature Rapid work hardening during machining Presence of hard abrasive carbide particles d. Generally low thermal conductivities and e. Tendency of chips to weld to cutting edges and form built-up edges.

Due to their high temperature strength, Inconel 718 remains hard and stiff at the cutting temperature, resulting in high cutting forces that promote chipping or deformation of the tool cutting edge. In addition, since superalloys retain a large percentage of their strength at elevated temperatures, more heat is generated in the shear zone resulting in greater tool wear than with most metals. Since the forces required to cut superalloys are about twice those required for alloy steels, tool geometry, tool strength, and rigidity are all important variables.

Their low thermal conductivities cause high temperatures during machining. The combination of high strength, toughness, and ductility impairs chip segmentation, while the presence of abrasive carbide particles accelerates tool wear. Inconel 718 also have a tendency to rapidly work harden which can create a

26 hardened surface layer that degrades the surface integrity and can lead to lower fatigue life. General guidelines for machining superalloys are very similar to those for titanium alloys [48].

2.6.1 Inconel 718 physical and mechanical properties

The workpiece chosen for this study was Inconel 718. It is a precipitationhardenable nickel-chromium alloy containing significant amounts of iron, niobium, and molybdenum along with lesser amount of aluminum and titanium (Table 2.2). Its combines corrosion resistance and high strength with outstanding weldability, including resistance to postweld cracking [46].

The alloy has creep-rupture strength at high temperatures up to 700C. Used in gas turbines, rocket motors, spacecraft, nuclear reactors, pumps and tooling [47]. The properties of Inconel 718 are listed in the Tables 2.3 and 2.4.

Table 2.2: Chemical properties of Inconel 718 (%) Element Ni (+Co) Fe Mo Ti C Si Cu Cr Co Nb (+Ta) Al Mn B Percentage (%) 50 - 55 Bal 2.3 3.3 0.65 1.15 0.08 0.35 0.3 17 21 1 4.75 5.5 0.2 0.8 0.35 0.006

27 Table 2.3: Physical properties of Inconel 718 Physical Properties Density Melting Point Range Specific Heat Average Coefficient of Thermal Expansion Thermal Conductivity Electrical Resistivity Curie Temperature Value 8.19 g/cm3 1260 - 1336C 435 J/kg.K 13.0 m/m.K 11.4 W/m.K 1250 n.m -112C

Table 2.4: Mechanical properties of Inconel 718 Mechanical Properties (Room Temperature) Ultimate Tensile Strength Yield Strength Eleongation (in 50mm) Elastic Modulus Value 1240 MPa 1036 MPa 12% 211 Gpa

2.7

Design of Experiment (DOE)

According to Lochner, R.H., and Matar, J.E. [49], design of experiment (DOE) is series of tests in which purposeful changes are made to the input variables of a process or system so that the reasons for change in the output responses can be observed and identified.

There are several reasons for designing complete factorial experiments rather than, for example, using a series of experiments investigating one factor at a time. The first is that factorial experiments are much more efficient for estimating main effects, which are the averaged effects of a single factor over all units. The second and very important reason is that interaction among factors can be assessed in a factorial experiment but not from series of one-at-a-time experiment. Interaction

28 effects are important in determining how the conclusions of the experiment might apply more generally. Complete factorial systems are often large, especially if an appreciable number of factors are to be tested. Often an initial experiment will set each factor at just two levels, so that important main effects and interactions can be quickly identified and explored further.

The choice of factors and the choice of levels for each factor are crucial aspects of the design of any factorial experiment, and will be dictated by the subject matter knowledge and constraints of time or cost on the experiment. The levels of factors can be qualitative or quantitative. The range of values for quantitative factor must be decide on how they are going to be measured and the level at which they will be control during the trials. Meanwhile, the quantitative factor is parameters that will be determine discretely [50].

2.7.1 Two-level full factorial design

Experiments with large numbers of factors are often used as a screening device to assess quickly important main effects and interactions. For this, it is common to set each factor at just two levels, aiming to keep the size of the experiment manageable. The levels of each factor are conventionally called low and high, or absent and present [50]. One of the advantages when implemented full factorial design is that it offers the capability to estimate the correlation between two or more factors in one time, where it is possible with other quality tool. Furthermore, this tool also capable to identify the importance factor in the experiment under a wide range of condition without sacrifices any factors.

Two-factor experiment is the simplest type of factorial design in DOE, in which effects of two factors on one or more response variables are tested simultaneously. It is common to use two levels for each factor studied, where k, is the number of factors and 2 indicates the level of experiment, then the total number

29 of combination will be 2k. In this experiment, four-factor experiments design was employed with two level of full factorial design experiment. Experiment shall include all the possible combination factors at two levels (low and high value of the machining parameters). The arrangement of the entire factor will be based on Design Expert version 7.0.0 software. This program will randomly and automatically analyze all the combination of the possible combination for the experiment. This software also can automatically analyze all the experimental results in order to investigate the influence of the machining parameters to the machining outputs or responses.

30

CHAPTER THREE

METHODOLOGY

3.1

Introduction

The purpose of this project is to evaluate the performance of WEDM on Inconel 718. To achieve this objective proper experimental plan is necessary to achieve good results. This experiment consist of four main elements namely, research design and data analysis, variables, research procedure and instrumentation. Figure 3.1 shows the four elements involved in research methodology in achieving the objective of this experiment.

Design of experiment with full factorial using Design Expert version 7.0.0 software was applied as a tool for design of experiment and data analysis. The confirmation test was also implemented in order to give the reliability of the WEDM results for Inconel 718.

31

Figure 3.1: Research methodology

3.2

Research Design and Data Analysis

Full factorial DOE will be employed for the whole design and analysis. DOE includes determining controllable factors and the levels to be investigated. In this experiment, four-factor experiment design will be use with two levels of full factorial design experiment. Based on this, the total number of experiments (combinations) required was 16 (24) experiment.

This experiment design includes all the possible combination factors at two levels (low and high value) for each parameter. Table 3.1 showed the notation used to denoted these levels; plus (+) for high value and minus (-) for low value. The arrangement of the factors for this project was based on Design Expert version 7.0.0 software. This program randomly choose the combination of factors to run the experiment and automatically analyse all the experimental results to investigate the influence of WEDM machining parameter on the surface integrity of Inconel 718.

32 Table 3.1: Two-level full factorial experiment with four factors. Factor Peak Current Pulse Duration (IP) (O ) ( s) (A) + + + + + + + + cp cp cp cp ( s) + + + + + + + + cp cp cp cp

Exp. o. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

Servo Voltage (SV) (V) + + + + + + + + cp cp cp cp

Pulse Interval (OFF) ( s) + + + + + + + + cp cp cp cp

otes: + (high value), - (low value), cp (centre point).

33 3.3 Research Design Variable

The design variables are described into two main groups, which are response parameters and machining parameters. Response parameters (machining characteristic @ dependent variables) include: a. b. c. d. Surface roughness, Ra Sparking gap, Gap Material removal rate, MRR Cutting speed, CS

Machining parameters or also known as independent variables involves in this experiment: a. b. c. d. Pulse duration (ON, s) Pulse interval (OFF, s) Peak current (IP, ampere) Servo voltage (SV, V)

ote: these parameters were donated as O , OFF, IP, SV respectively.

3.3.1

Machining parameters

Based on previous studies, several numbers of machining factors have been used in WEDM operation. As mentioned earlier, electrical parameters are the factor that significantly influences the machining characteristic whereby, non-electrical parameters have less significant to the machining characteristic [25 - 27].

34 Most researchers identified four WEDM cutting parameters that greatly affect the machining output; ON, OFF, IP, SV which are known as electrical factors while nonelectrical factor (mechanical factor) include wire tension, wire speed, dielectric pressure etc are held constant. Table 3.2 shows the setting of the parameters studies. All the suggested value in Table 3.2 was based on previous studies. The actual values of these setting are shown in Table 3.3

Table 3.2: The design of machining parameters Level Machining Parameters Pulse Duration ON ( s) Pulse Interval OFF ( s) Peak Current IP (Ampere) Servo Voltage (Volt) 1 (Low) Code Value 001 007 2210 030 True Value 0.65 4.0 8.0 30.0 2 (High) Code True Value Value 003 0.75 015 2215 060 8.0 12.0 60.0

35 Table 3.3: Actual value of experiment design Machining Voltage Wire Speed Wire Tension Injection Pressure : : : : SV Exp. o. Servo Voltage (V) 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 30 60 30 60 30 60 30 60 30 60 30 60 30 60 30 60 45 45 45 45 80V 10 m/min 800 g 12 bar IP Peak Current (A) 8 8 12 12 8 8 12 12 8 8 12 12 8 8 12 12 10 10 10 10 O Pulse Duration ( s) 0.65 0.65 0.65 0.65 0.75 0.75 0.75 0.75 0.65 0.65 0.65 0.65 0.75 0.75 0.75 0.75 0.70 0.70 0.70 0.70 OFF Pulse Interval ( s) 4 4 4 4 4 4 4 4 8 8 8 8 8 8 8 8 6 6 6 6

36 3.3.2 Machining characteristic

This study investigates the machining characteristics such as surface roughness (Ra), spark gap (Gap), material removal rate (MRR) and cutting speed (CS). These are the most common key indicators used by many manufacturers to determine the quality machine surface through surface roughness, while spark gap is a reflection of degree of accuracy the WEDM machining can achieved and material removal rate with the input of cutting speed is the key indicator of the efficiency of the WEDMing process [31].

3.3.3

Surface roughness, Ra

Surface topography or surface roughness, also known as surface texture are terms used to express the general quality of a machined surface, which is concerned with the geometric irregularities and the quality of a surface [9]. According to Armarego and Brown [44], ideal surface roughness may be specified in a variety of ways, but two common methods are the peak to valley height (h) and the arithmetic average, Ra (m).

The Ra value, also known as centre line average (CLA) and arithmetic average (AA) is obtained by averaging the height of the surface above and below the centre line. The Ra will be measured using a surface roughness tester from Mitutoyo, Model: Formtracer CS-5000. Before conducting the measurement, all the samples were cleaned with acetone. The Ra values of the WEDMed surface were obtained by averaging the surface roughness values of 5mm measurement length.

37 3.3.4 Sparking gap, Gap

Sparking gap (SG) or also known as overcut is one of the responses investigate in this study. Sparking gap is measure using optical microscope in order to study the correlation between machining parameters and the spark gap. The unit used is mm. Sparking gap (Gap) can be calculated by the following formula:

Spark Gap =

(kerf width wire diameter) 2

3.3.5

Cutting speed, CS

Cutting speed (CS) is measured after WEDMing 10mm distance and recorded using the WEDM machine controller.

3.3.6

Material Removal Rate, MRR

The material removal rate (MRR) of the workpiece is the volume of the material removal per minute. As for these the following equation is used to determine the material removal rate (MRR) value:

38 Volume = Spark Gap (mm) x Machine Distance (10mm) x Workpiece Thickness (25mm) By knowing the density of Inconel 718 was 0.00819 g/mm3.The mass of material that removed by the WEDM process:

Mass = Density (g/mm3 ) x Volume (mm3)

Therefore, MRR is then measured by:

MRR =

Mass (g) Machining Time (min)

3.4

Research Procedure

Based on previous studies, the following four machining parameters i.e. pulse ontime, pulse off-time, peak current and servo voltage were chosen as the input parameters. WEDM performance on Inconel 718 is measured by four important response parameters such as surface finish (Ra), sparking gap (Gap), material removal rate (MRR) and cutting speed (CS).

Full factorial design was employed with two level of full factorial design experiment. The total number of experiment (combinations) required is 24 = 16 experiments with additional of 4 centre point. The experiment design will include all the combination factors at two levels. The arrangement of the factor for this project will be

39 based on design expert version 7.0.0 software. Table 3.4 shows the flowchart of the step involves in the overall experiment.

40

START

Design Plan (Full Factorial Design)

20 experiment including 4 centre points

Run experiment according to design plan

Experiment resulted obtained

Four responses: surface finish, spark gap, MRR & cutting speed

Performed analysis based on full factorial design

ANOVA

Determine the optimum machining condition for the work material

Confirmation test

To verify the experimental results

Final results recorded

E D

Figure 3.2: Flowchart of experiment stepping

41 3.5 Experimental Set-up

The experiments were performed on a WEDM linear motor 5 axes Sodick series AQ537L machine. Zinc coated brass wire with diameter of 0.25mm was chosen as electrode in machining Inconel 718. Table 3.3 indicates the list of machining parameters that involve in this experiment. These parameters were kept constant throughout the experiment trials.

Table 3.4: Machining parameters set-up (constant parameters) Parameter Main Power Supply Voltage, V (Volt) Servo Speed, SF (mm/mmin) Wire Tension, WT (g) Wire Speed, WS (m/min) Flushing Pressure, FP (bar) Wire Electrode Polarity Dielectric Fluid Setting Value 80 At no load) normal servo control 800 10 12 Zinc coated brass wire, 0.25mm Workpiece : Negative Wire Electrode : Positive Submerge deionizer water.

The size of Inconel 718 was a rectangular bar with dimension of 48mm x 44mm x 25mm. The workpice material was cut to size for 10mm length with a 3mm gap between two cutting experiments. Permanent marker will be used to mark the cut out workpiece to indicate the orientation accordingly to the parameter. Usage of scriber to mark the workpiece need to be avoided to preventing coated being snapped when passes the scriber mark. The wire is so sensitive that any irregularities on the machine surface can cause it to break.

42 To measure the surface roughness, 5.0mm will be cut from the workpiece and the remaining is left to the remaining workpiece as a backup sample and a tool to measure the spark gap.

The spark gap was measure using Zeiss Axiotech High Power Optical Microscope with 100x magnification. All the measurement of spark gap was measured before the original workpiece was cut into smaller size of 3mm x 5mm x 25mm. This is required in order to facilitate post measurements on other equipment such as SEM and surface roughness tester.

After the machining trials were completed, the machined workpiece were cut out into smaller specimens perpendicular to the cutting surface with the same machine. This is to reveal the section of machined surface layer for measuring the surface roughness. The machine surface roughness was the measured using Mitutoyo surface roughness machine. In accordance to the research made on measuring the surface roughness, Mustafa et al. [45] reported that the measurement must consider x & y direction perpendicular to the lay direction. Horizontal direction (x-direction) is expected to be more crucial than the averages surface roughness along the vertical direction (ydirection) due to the position of lay direction, the average of surface roughness along horizontal and vertical direction is used as an indication of the total surface roughness of each test section as shown below.

Roughness on X-Axis (Rx) = (RX1 + RX2 + RX3)/3 Roughness on Y-Axis (Ry) = (RY1 + RY2 + RY3)/3 Overall Roughness (RAV) = (Rx + RY)/2

43 The method was followed since other published papers available do not provide any specific information on the selection of machining parameters for various machining conditions and materials.

3.6

Experimental Equipments

The equipment involved in this study are as follows: a. WEDM machine WEDM linear motor 5 axes Sodick series AQ537L (Figure 3.3). b. Measuring equipment surface roughness was measured with the Mitutoyo surface roughness measuring machine (Figure 3.4). Whereby, Zeiss Axiotech High Power Optical Microscope will used to measure the spark gap (Figure 3.5).

44

Figure 3.3: WEDM linear motor 5 axes Sodick series AQ537L

Figure 3.4: Mitutoyo surface roughness measuring machine

45

Figure 3.5: Zeiss Axiotech high power optical microscope

46

CHAPTER FOUR

EXPERIME TAL RESULTS A D DATA A ALYSIS

4.1

Introduction

This chapter discusses on the experimental results on WEDM of Inconel 718 using zinc coated brass wire of diameter 0.25mm. The main purpose of this research is to investigate the performance of the WEDM on Inconel 718 based on predetermined WEDM machining parameters such as servo voltage (SV), peak current (IP), Pulse Duration (ON) and Pulse Interval (OFF).

Design Expert software version 7.0.0 was employed to analyse all the data of the 20 experiment trial runs by using ANOVA approach. The quadratic mathematical models were proposed for the response variables such as surface roughness (Ra), sparking gap (GAP), material removal rate (MRR) and cutting speed (CS). These relationships between machining factors and responses were evaluated by the F-test of ANOVA and the fit summary reveals that the fitted quadratic model is statistically significant to be considered.

47 4.2 Experimental Results

Full factorial design of four factors with two levels each was conducted which consist of 20 runs (including of four center points). The machine responses were surface roughness (Ra), sparking gap (Gap), material removal rates (MRR) and cutting speed (CS) respectively. Tables 4.1 to 4.4 show the summary of the machining responses corresponding to the various setting of WEDM machine parameters.

Table 4.1 shows the summary of surface roughness measurement of the experimental trials. Taylor Mitutoyo surface roughness measuring machine was used to conduct all the surface roughness value. The measurement length of each specimen was 5mm and it was divided into 3 sections with sampling of 0.25mm each. Every section was measured 3 times before average results of each section were obtained. Table 4.1 shows the summary of the measurement obtain for the surface roughness measurement.

48 Table 4.1: Experimental results of surface roughness No. of trial 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. Rx ( m) 2.13 1.79 2.11 2.29 2.64 2.66 2.59 2.74 1.96 1.94 2.06 2.11 2.62 2.51 3.01 2.36 2.19 2.16 2.24 2.56 Ry ( m) 2.09 1.68 1.68 2.06 2.54 2.23 2.07 2.54 1.91 2.01 1.93 1.89 2.57 2.10 2.42 1.91 1.73 1.98 2.41 2.06 Total Average ( m) 2.11 1.74 1.90 2.18 2.59 2.45 2.33 2.64 1.94 1.98 2.00 2.00 2.60 2.31 2.72 2.14 1.96 2.07 2.33 2.31

Table 4.2 shows the summary of kerf width and sparking gap of the experimental runs. All measurement was taken using Zeiss Axiotech High Optical Microscope under 100 times of magnification. In order to reduce the uncertainty errors, the kerf width measurement was taken three times at three different points along to the cutting line. This measurement was also taken on the top and the bottom of the cutting line before the average measurement value of kerf width was calculated. The sparking gap was calculated using the following equation:

49

Where, Kerf width was obtained from the measurement shown on Appendix D. Wire diameter = 0.25mm

Table 4.2: Experimental results of sparking gap (Gap) No. of trial Sparking gap on top surface (mm) 0.031 0.031 0.032 0.039 0.043 0.038 0.039 0.039 0.036 0.034 0.035 0.034 0.036 0.037 0.042 0.047 0.037 0.036 0.032 0.032 Sparking gap on bottom surface (mm) 0.024 0.026 0.026 0.031 0.038 0.042 0.036 0.039 0.029 0.026 0.028 0.025 0.037 0.041 0.039 0.029 0.016 0.025 0.026 0.027 Total Average ( m) 0.028 0.029 0.029 0.035 0.040 0.040 0.037 0.039 0.032 0.030 0.031 0.029 0.036 0.039 0.041 0.038 0.027 0.030 0.029 0.030

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

50 Other machine response that has been considered was cutting speed (CS). This response was recorded by the WEDM machine time indicator. The distance of 10mm for cutting distance was fixed for all experiment. Cutting speed measurement value showed in Table 4.3 is obtained with following equation:

Where, Machining distance, d = 10mm Machining time, t was obtained from the WEDM machine time indicator

Table 4.3: Experimental results for cutting speed (CS) No. of trial Machining distance (mm) 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 Machining time (min) 13.65 9.57 12.95 8.78 8.85 6.57 9.00 6.23 11.93 10.13 11.77 9.47 8.28 7.15 8.53 6.85 7.18 7.25 7.20 7.30 Cutting speed, CS (mm/min) 0.733 1.045 0.772 1.139 1.130 1.523 1.111 1.604 0.838 0.987 0.850 1.056 1.207 1.399 1.172 1.460 1.392 1.379 1.389 1.370

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

51

Meanwhile, Table 4.3 indicated the experimental results for material removal rates (MRR) for each 20 run of experiments. MRR shown on Table 4.5 were determined by using the following equation:

Where, Mass is defined by:

Density of the Inconel 718 = 0.00819 g/mm3 Volume is definied by:

Volume = sparking gap, SG (mm) x machining distance (10mm) x thickness, t (25mm)

Machining time, t was recorded by the WEDM machine time indicator.

52 Table 4.5: Experimental results for material removal rate (MRR) No. of trial Sparking gap (mm) 0.028 0.029 0.029 0.035 0.040 0.040 0.037 0.039 0.032 0.030 0.031 0.029 0.036 0.039 0.041 0.038 0.027 0.030 0.029 0.030 Volume (mm3) 7.000 7.250 7.250 8.750 10.000 10.000 9.250 9.750 8.000 7.500 7.750 7.250 9.000 9.750 10.250 9.500 6.750 7.500 7.250 7.500 Mass Machining time (min) 13.650 9.567 12.950 8.783 8.850 6.567 9.000 6.233 11.933 10.133 11.767 9.467 8.283 7.150 8.533 6.850 7.183 7.250 7.200 7.300 MRR (g/mm3) 0.0042 0.0062 0.0050 0.0082 0.0093 0.0125 0.0084 0.0128 0.0055 0.0060 0.0054 0.0062 0.0089 0.0112 0.0098 0.0114 0.0077 0.0084 0.0082 0.0084

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

(g) 0.057 0.059 0.059 0.072 0.082 0.082 0.076 0.080 0.066 0.061 0.063 0.059 0.074 0.080 0.084 0.078 0.055 0.061 0.059 0.061

Table 4.6 present the overall results for wire electrical discharge machining on Inconel 718 in terms of surface roughness (Ra), sparking gap (Gap), material removal rates (MRR) and cutting speed (CS). All these machining responses were used as input to the Design Expert version 7.0.0 software for further analysis. By using ANOVA, information such as main effects, percentage contribution for each factor and estimation of the optimum results can be produced and analysed.

53 Table 4.6: Overall experimental results corresponded to each run Exp. No. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. Factors IP ON (A) ( s) 8 0.65 8 12 12 8 8 12 12 8 8 12 12 8 8 12 12 10 10 10 10 0.65 0.65 0.65 0.75 0.75 0.75 0.75 0.65 0.65 0.65 0.65 0.75 0.75 0.75 0.75 0.7 0.7 0.7 0.7 Responses Gap MRR (mm) (g/min) 0.028 0.0042 0.029 0.029 0.035 0.040 0.040 0.037 0.039 0.032 0.030 0.031 0.029 0.036 0.039 0.041 0.038 0.027 0.030 0.029 0.030 0.0062 0.0050 0.0082 0.0093 0.0125 0.0084 0.0128 0.0055 0.0060 0.0054 0.0062 0.0089 0.0112 0.0098 0.0114 0.0077 0.0084 0.0082 0.0084

SV (V) 30 60 30 60 30 60 30 60 30 60 30 60 30 60 30 60 45 45 45 45

OFF ( s) 4 4 4 4 4 4 4 4 8 8 8 8 8 8 8 8 6 6 6 6

Ra ( m) 2.11 1.74 1.90 2.18 2.59 2.45 2.33 2.64 1.94 1.98 2.00 2.00 2.60 2.31 2.72 2.14 1.96 2.07 2.33 2.31

CS (mm/min) 0.733 1.045 0.772 1.139 1.130 1.523 1.111 1.604 0.838 0.987 0.850 1.056 1.207 1.399 1.172 1.460 1.392 1.379 1.389 1.370

54 4.3 Analysis of Results

This section discusses the experimental finding of parametric influences on the performance characteristics of WEDM machined on Inconel 718. The analysis of variance (ANOVA) was used to generate statistically the significant machining parameters and the percentage contribution of each parameter. As mention earlier, Design Expert version 7.0.0 software was used to analyze the ANOVA analysis. ANOVA table is commonly used to summarize the experimental results. This table concludes all information of analysis of variance and case statistics for further interpretation [9].

In the next section, all the analyses were presented in normal probability plot, main effect plot and interaction plot for the dependent parameters that significant to the responses. The interpretation was done unilaterally, meaning that ANOVA analysis for all 4 responses was done separately at one time.

4.3.1 Analysis results for surface roughness (Ra)

According to the analysis done by the Design Expert software, if the values of probability (Prob>F) are less than 0.05, it indicated that the factors is significant to the response parameters. As observed in ANOVA result (Table 4.7), there is only one factor that influences surface roughness (Ra). In this case, the pulse on (O ) were significant to the Ra. Other factors, namely servo voltage (V), peak current (IP), pulse off (OFF) are not significant since the probability values are greater than 0.1, therefore, these factor did not appear on Table 4.7. In this investigation, 95% of confidence interval (CI) was used.

55 The lack of fit was not significant which satisfy the model to be fitted. The value of R2 was quiet high and closed to 1 ( 0.6589) which is desirable. The adjusted R2 and predicted R2 were in agreement as the difference between the values was below 0.2 ( 0.1126). The adequate prediction is above value of 4 ( 7.398), thus indicated that the model discrimination was adequate.

Table 4.7: ANOVA for surface roughness (Ra)

The information was better illustrated in normal probability plot as shown in Figure 4.1. Normal probability plot is needed in order to check for the normality of residuals of the factors studied.

Figure 4.1 reveals that residuals are spread on a straight line implying that errors are distributed normally. Meanwhile, the plot in Figure 4.2 shows no obvious pattern

56 and unusual structure and all the results fall in the acceptance range. Therefore, it can be concluded that the model proposed was adequate. For clearer observation, the main plot in Figure 4.3 indicates how the significant variables affected the Ra.

Figure 4.1: Normal probability plot of residuals for surface roughness (Ra)

57

Figure 4.2: Residual vs. predicted response for surface roughness (Ra)

2.4725

1.9813

Figure 4.3: Main effect plot for surface roughness (Ra)

58 As shown in Figure 4.4, the factor that influence the Ra is pulse on (ON) and its clearly indicated that whenever ON is increased from 0.65 s up to 0.75 s, the value of Ra is increased dramatically. The percentage of the increment of Ra was approximately 24.8%. From the correlation obtained in this investigation, the judgement in terms of selecting the most suitable setting for ON for future optimization can be made. In order to obtain better Ra during WEDM of Inconel 718, ON should be set at the lowest value, 0.65 s.

The mathematical model for Ra was also developed by the ANOVA analysis in order to identify the relationship between independent variables, namely, ON and Ra. The following equation is the final empirical models in terms of coded and factors and actual factors for Ra respectively. This equation was generated by the Design Expert software after the transformation had been carried out.

The final equation in terms of coded factors: Ra = +2.33 + 0.25*C

The final equation in terms of actual factors: Ra = -1.21187 + 4.91250*ON

4.3.2 Analysis results for sparking gap (Gap)

Similar approach is also employed to the next response, sparking gap (Gap). The final ANOVA table for sparking gap (Gap) is shown in Table 4.8

59 Table 4.8: ANOVA for sparking gap (Gap)

In terms of sparking gap (Gap), it was observed that only one factor without any interaction have the major influence to the sparking gap, its also the factor of pulse on (O ). The value of probability for ON effect was below 0.05 with confidence interval (CI) is 95%. The lack of fit was not significant which satisfy the model to be fitted. The value of R2 was high and closed to 1 ( 0.8025) which is desirable. The adjusted R2 and predicted R2 were in agreement as the difference between the values was below 0.2 ( 0.0019). The adequate prediction value was above 4 ( 13.249), thus indicated that the model discrimination was adequate.

60

Figure 4.4: Normal probability plot of residuals for sparking gap (Gap)

61

Figure 4.5: Residual vs. predicted response for sparking gap (Gap)

0.0386

0.0304

Figure 4.6: Main effect plot for sparking gap (Gap)

62 Figure 4.5 reveals that residuals are spread on a straight line implying that errors are distributed normally. In other hand, the plot in Figure 4.6 shows no obvious pattern and unusual structure and all the results fall in the acceptance range. Accept for the run number 13 which lies far from other runs number, however it still fall in the acceptable range. Therefore, it can be concluded that the model proposed was adequate. For clearer observation, the main plot in Figure 4.7 indicates how the significant variables affected the Gap.

As shown in Figure 4.7, increasing in pulse on (ON) led to an increase of sparking gap (Gap). Once again, ON revealed an interesting plot with an increment of 27% as it increased from 0.65 s up to 0.75 s. Previous studies reported that, it is expected that sparking gap continues to increase if the range of ON is widen. Based on the graph, the judgement in term of selecting the most suitable setting for ON for future optimization can be made. In order to obtain better Gap during WEDM of Inconel 718, ON should be set at the lowest value of 0.65 s. The following equations are the final empirical models in terms of coded and factors and actual factors for Gap respectively.

The final equation in terms of coded factors: Gap = +0.035 + 4.18 x 10-3*C

The final equation in terms of actual factors: Gap = -0.024062 + 0.083750*ON

63 4.3.3 Analysis results for material removal rate (MRR)

Material removal rate (MRR) in WEDM processes is an important factor because of its vital effect on the production cost [4]. Table 4.9 indicates the final analysis of ANOVA for material removal rate (MRR).

Table 4.9: ANOVA for material removal rate (MRR)

By considering the results from the ANOVA for MRR, there were three main significant main effects that influence the MRR. ANOVA of MRR also indicated one interaction of main effect for the MRR. The significant main effects and interaction were identified by the probability value (Prob>F) justification. With the confidence interval

64 (CI) of 95%, whatever main effect or interaction with their Prob>F value of 0.05 or lower, the main effect are considered as the significant factors that affecting MRR.

Results from ANOVA (Table 4.9) showed the main significant factors were servo voltage (SV), pulse on (ON) and pulse off (OFF). Meanwhile, interaction between servo voltage and pulse off (SV*OFF) were observed to be the significant (Prob>F 0.0076) interaction in this study.

The lack of fit was not significant which satisfy the model to be fitted. The value of R2 was high and closed to 1 ( 0.9558) which is desirable. The adjusted R2 and predicted R2 were in agreement as the difference between the values was below 0.2 ( 0.0358). The adequate prediction value was above 4 ( 23.661), thus indicated that the model discrimination was adequate.

Figure 4.8 reveals that residuals are spread on a straight line implying that errors are distributed normally. In other hand, the plot in Figure 4.9 show no obvious pattern and unusual structure and all the results fall in the acceptance range. Accept for the run number 4 which lies far from other runs number, however it still fall in the acceptable range. Therefore, it can be concluded that the model proposed was adequate. For clearer observation, the main plots in Figures 4.8 to 4.11 indicate how the significant variables affected the MRR.

65

Figure 4.7: Normal probability plot of residuals for material removal rate (MRR)

Figure 4.8: Residual vs. predicted response for material removal rate (MRR)

66

0.0105 0.0093

0.0071 0.0058

0.0083

0.0081

Figure 4.9: Main effect plot for material removal rate (MRR)

From the main plot graph shown in Figure 4.10, it was obvious that an increase in both SV and ON factors leads to the increase of the MRR. As observed in Figure 4.10, MRR experienced an increment percentage of approximately 31% when SV is increase from 30V up to 60V. Similar pattern was revealed when ON was increased from 0.65 s to 0.75 s, the MRR increases with huge percentage of approximately 81%. Apparently, opposite result is revealed from the pulse off (OFF) factor. From the graph shown in Figure 4.10, when OFF increases from 4 s up to 8 s, the MRR slightly decreases about 2.4% respectively. Based on these relationships, maximum MRR can be obtained when the parameters are set at SV = 60V, ONN = 0.75 s and OFF = 4 s.

67

SV = 60V OFF = 4 s MRR = 0.0099 SV = 30V OFF = 8 s MRR = 0.0074 SV = 60V OFF = 8 s MRR = 0.0087

SV = 30V OFF = 4 s MRR = 0.0067

Figure 4.10: Interaction between SV*OFF for material removal rate (MRR)

Figure 4.11 shows an interaction of the factors graphically for MRR. It was observed when SV and OFF were set at 60V and 8 s respectively the MRR increase about 17.6%. Meanwhile, MRR increases more rapidly to 47.8% with an increment of SV as OFF were set at 4 s. In order to gain maximum MRR the parameters of SV and OFF should be set up to 60V and 4 s respectively. The following equations is the final empirical models in terms of coded factors and actual factors for MRR respectively.

The final equation in terms of coded factors: MRR = +8.187 x 10-3 + 1.125 x 10-3*A + 2.350 x 10-3*C 1.375 x 10-4*D 4.750 x 10-4*A*D

The final equation in terms of actual factors: MRR = -0.031950 + 1.7 x 10-4*SV + 0.047*ON + 6.43750 x 10-4*OFF - 1.58333 x 10-5*SV*OFF

68 4.3.4 Analysis results for cutting speed (CS)

Lastly, the final analysis in this chapter is to determine the factors and interaction that affect the cutting speed (CS). Similar procedures were applied for the other responses; the significant factors and the possible interaction of CS are referred to the ANOVA result shows in Table 4.10. Whatever factors that have the Prob>F less than 0.05 are considered as significant factor for CS with the confident interval (CI) used is 95%.

Table 4.10: ANOVA for cutting speed

69 Based on ANOVA in Table 4.10, results show that there are five effects with Prob>F value below 0.05, which are significant for a 95% of confident interval (CI). These are servo voltage (SV), peak current (IP), pulse on (ON), the interaction between servo voltage (SV) and peak current (IP) and finally, the interaction between servo voltage (SV) and pulse off (OFF). Meanwhile from the ANOVA table, factor OFF with probability value of 0.7386 (> 0.05) is shown because it is required in order to support the hierarchy in the software and this factor was not discussed any further. Apparently, the analysis for CS was quite complicated since the number of effects and interactions were higher than the previous responses.

The lack of fit was not significant which satisfy the model to be fitted. The value of R was high and closed to 1 ( 0.9949) which is desirable. The adjusted R2 and predicted R2 were in agreement as the difference between the values was below 0.2 ( 0.0054). The adequate prediction value was above 4 ( 62.340) as obtained in this case, thus indicated that the model discrimination was adequate.
2

Figure 4.12 showed that residuals are spread on a straight line implying that errors are distributed normally. In other hand, the plot in Figure 4.13 shows no obvious pattern and unusual structure and all the results fall in the acceptance range. Accept for runs number 4 and 17 which lies far from other runs number, however they still fall in the acceptable range. Therefore, it can be concluded that the model proposed was adequate. For clearer observation, the main plots in Figures 4.12 to 4.15 indicate how the significant variables affected the CS.

70

Figure 4.11: Normal probability plot of residuals for cutting speed (CS)

Figure 4.12: Normal probability plot of residuals for cutting speed (CS)

71

1.2663 1.1307

1.0936 0.9679

1.3196

0.9222

Figure 4.13: Main effect plot for cutting speed (CS)

The final ANOVA results for CS obtained from the modified model are shown in Table 4.10. Results show that there are three main effects with Prob>F value below 0.05, which are significant for a 95 confident interval (CI) level. As displayed in the Figure 4.14, CS was affected by SV, IP and ON. For SV factor, increasing in SV from 30 to 60V increases the CS up to 30.8%. Meanwhile, CS was observed to slightly increase about 3.4% by the increased of the IP from 8A up to 12A. Lastly, as shown graphically in Figure 4.14, ON shows the significant factor for CS until up to 40.1% of total increasing in CS value. Based on these relationships, maximum CS can be obtained when the parameters are set at SV = 60V, IP = 12A and ON = 0.75 s.

72

SV = 60V IP = 12A CS = 1.3051 SV = 30V IP = 8A CS = 0.9669 SV = 60V IP = 8A CS = 1.2280

(a) Interaction between SV*IP

SV = 60V OFF = 4 s CS = 1.3268 SV = 30V OFF = 8 s CS = 1.0093 SV = 60V OFF = 8 s CS = 1.2167 SV = 30V OFF = 4 s CS = 0.9273

(b) Interaction between SV*OFF

Figure 4.14: Interaction plot of cutting speed (CS)

73 Figures 4.15 (a) and (b) show the significant interaction factors of the parameters for the CS. Based on the ANOVA results (Table 4.10), the Prob>F value of SV*IP, SV*OFF interaction were 0.0057 and < 0.0001 respectively. As the value of Prob>F mentioned is closer to zero, it indicates that the model of interaction factor was significant. Therefore, the first interaction to be considered was the interaction between servo voltage and peak current (SV*IP) as shown graphically on Figure 4.15 (a). When IP was set at 8A and SV was vary from 30V to 60V, CS increased about 27%. On the other hand, CS percentage increased about 35% when IP was set at low level, 4 s with SV vary remain unchanged. This 35% increment for CS is obtain when SV = 60V and IP = 12A.

Next interaction was the interaction between servo voltage and pulse off (SV*OFF).as shown in Figure 4.15 (b). When SV is set between 30V and 60V with a constant OFF of 8 s, the CS experienced a 20.6% increment. Similar pattern was recorded as OFF is set at low value of 4 s with the same setting of SV, it contributed about 43.1% of increment in CS value. These results implied that in order to achieved maximum value of CS the SV and OFF should be set at 60V and 4 s respectively. The following equations are the final empirical models in terms of coded factors and actual factors for CS respectively.

The final equation in terms of coded factors: Sqrt(CS) = +1.05 + 0.071*A + 8.814 x 10-3*B + 0.094*C - 6.955 x 10-4*D + 8.315 x 10-3*A*B - 0.022*A*D

The final equation in terms of actual factors: Sqrt(CS) = -0.58760 + 6.25238 x 10-3*SV - 8.06548 x 10-3*IP + 1.88401*ON + 0.031963*OFF + 2.77165 x 10-4*SV*IP - 7.18012 x 10-4*SV*OFF

74 Summary of the significant factors that were achieved from WEDM experimental run is shown in Table 4.11. These include all the responses investigated in this experiment.

Table 4.11: Summary of the significant factors in WEDM Inconel 718 Responses Surface roughness (Ra) Sparking gap (Gap) Material removal rate (MRR) Cutting speed (CS) Significant Factor C C A-C-D-AD A-B-C-D-AB-AD

ote: A = Servo voltage, B = Peak current, C = Pulse on, D = Pulse off

4.4

Confirmation Test

The confirmation test is the final step undertaken during this experiment. The purpose of the confirmation runs is to validate the conclusion drawn during the analysis phases [35]. In addition, the confirmation tests need to be carried out in order to ensure that the theoretical predicted model for optimum results using the software was accepted or in other word to verify the adequacy of the models that were developed. All parameters used in the confirmation test were suggested by Design Expert software. Three (3) confirmation tests were carried out in order to compare the experimental results from the prediction made by the ANOVA. Table 4.12 shown in this section indicates the optimization of quality characteristic needed for each response.

75 Table 4.12: Quality characteristic of the machining performance Machining Characteristic Surface roughness (Ra) Sparking gap (Gap) Material removal rate (MRR) Cutting speed (CS) Quality Characteristic Minimum Minimum Maximum Maximum

4.4.1 Confirmation tests and results

Table 4.13 shows the three series of parameters settings for the confirmation test. The parameters values were selected between the high and low range of the machining factor that have been studied from previous experiment.

Table 4.13: True value of confirmation test experiment Machining Voltage Wire Speed Wire Tension Injection Pressure Exp. No. 1. 2. 3. : : : : SV Servo Voltage (V) 30 30 30 80V 10 m/min 800 g 12 bar IP Peak Current (A) 8 11.2 11.6

ON Pulse Duration ( s) 0.65 0.65 0.65

OFF Pulse Interval ( s) 4 5.5 7

Tables 4.14 to 4.17 show the results of the machining responses for surface finish (Ra), sparking gap (Gap), material removal rate (MRR) and cutting speed (CS) respectively.

76 Table 4.14: Confirmation test results for surface roughness (Ra) No. of trial 1. 2. 3. Rx ( m) 1.92 2.19 2.13 Ry ( m) 1.75 2.13 1.84 Total Average ( m) 1.84 2.16 1.99

Table 4.15: Confirmation test results for sparking gap (Gap) No. of trial Sparking gap on top surface (mm) 0.032 0.029 0.031 Sparking gap on bottom surface (mm) 0.022 0.026 0.031 Total Average ( m) 0.027 0.028 0.031

1. 2. 3.

Table 4.16: Confirmation test results for cutting speed (CS) No. of trial Machining distance (mm) 10 10 10 Machining time (min) 14.20 13.70 13.65 Cutting speed, CS (mm/min) 0.704 0.730 0.733

1. 2. 3.

Table 4.17: Confirmation test results for material removal rates No. of trial Sparking gap (mm) 0.027 0.028 0.031 Volume (mm3) 6.750 7.000 7.750 Mass Machining time (min) 14.20 13.70 13.65 MRR (g/mm3) 0.0039 0.0042 0.0047

1. 2. 3.

(g) 0.055 0.057 0.063

77 Full confirmation test measurement results for sparking gap (Gap) can be referred to Appendix D.

4.5

Comparison of the Test Results

Based on the flow charts of experiment steps discussed in chapter three, the comparison of the test results between the theoretically prediction and confirmation test results was the final consideration that will evaluate whether the optimum parameters predicted were in the allowable range. The margin of error from the prediction and experimental results was set below than 15%. Margin error was calculated using the equation below:

Tables 4.18 to 4.21 show the comparison of test results between theoretical prediction and confirmation test for surface roughness (Ra), sparking gap (Gap), material removal rate (MRR) and cutting speed (CS) respectively.

78 Table 4.18: Comparison test results for surface roughness (Ra) No. of confirmation run 1. 2. 3. Experimental (Confirmation test) 1.84 2.16 1.95 Prediction (Design Expert) 1.98 1.98 1.98 Error Margin (%) 7.68 8.33 1.54

Table 4.19: Comparison test results for sparking gap (Gap) No. of confirmation run 1. 2. 3. Experimental (Confirmation test) 0.027 0.028 0.031 Prediction (Design Expert) 0.030 0.030 0.030 Error Margin (%) 11.11 7.14 3.23

Table 4.20: Comparison test results for material removal rate (MRR) No. of confirmation run 1. 2. 3. Experimental (Confirmation test) 0.0039 0.0042 0.0047 Prediction (Design Expert) 0.0044 0.0046 0.0049 Error Margin (%) 12.82 9.52 4.26

Table 4.21: Comparison test results for cutting speed (CS) No. of confirmation run 1. 2. 3. Experimental (Confirmation test) 0.704 0.730 0.733 Prediction (Design Expert) 0.754 0.783 0.810 Error Margin (%) 7.10 7.26 10.51

79 4.6 Verification of the Mathematical Models

From ANOVA analysis, the mathematical models for each response were generated by the Design Expert software. These mathematical models identified the relationship between the independent variables (SV, IP, ON and OFF) and the dependent variables (Ra, Gap, MRR and CS). Even though the mathematical models are able to predict the results automatically when setting parameters are inserted into the system, these mathematical models still require verification.

This section verifies the mathematical models developed by the Design Expert software. Some examples of experimental data were selected and manually calculated using the equation. This is to make sure that the predicted results given by the software is correct. In this case, all calculation was based on the setting parameters used in the experimental data trial # 1 (SV = 30V, IP = 8A, ON = 0.65 s, OFF = 4.0 s). The mathematical models for all responses were presents as below:

a.

Surface roughness (Ra) From Table 4.7, the surface roughness (Ra) can be obtain from: Ra = +2.33 + 0.25(C) By using experimental data trial #1, the predicted Ra was calculated as follows; Ra = +2.33 + 0.25(C), where; C = -ve (low)

Ra = 2.33 + 0.25(-1) = 2.08 m #

From the experimental data trial #1 results, the value of Ra is 2.11 m. So the margin error is 1.42%.

80 b. Sparking Gap (Gap) Similarly from Table 4.8, the sparking gap (Gap) can be obtained from: Gap = +0.035 + 4.18 x 10-3(C) By using experimental data trial #1, the predicted Gap was calculated as follows; Ra = +0.035 + 4.18 x 10-3(C) where; C = -ve (low) Ra = +0.035 + 4.18 x 10-3(-1) = 0.0308mm #

From the using experimental data trial #1 results, the value of Gap is 0.0304 m. So the margin error is 1.32%.

c.

Material removal rate (MRR) Similarly from Table 4.9, material removal rate (MRR) can be obtained from: MRR = +8.187 x 10-3 + 1.125 x 10-3*A + 2.350 x 10-3*C 1.375 x 10-4*D 4.750 x 10-4*A*D By using experimental data trial #1, the predicted MRR was calculated as follows; MRR = +8.187 x 10-3 + 1.125 x 10-3(A) + 2.350 x 10-3(C) 1.375 x 10-4(D) 4.750 x 10-4(A)(D), where; A = -ve (low) B = -ve (low) C = -ve (low) D = -ve (low) MRR = +8.187 x 10-3 + 1.125 x 10-3(-1) + 2.350 x 10-3(-1) 1.375 x 10-4(-1) 4.750 x 10-4(-1)(-1) = 0.0044 g/min #

81 From the experimental data trial #1 results, the value of MRR is 0.0042 m. So the margin error is 4.76%.

d.

Cutting speed (CS) Similarly from Table 4.8, the cutting speed (CS) can be obtained from: Sqrt(CS) = +1.05 + 0.071(A) + 8.814 x 10-4(B) + 0.094(C) - 6.955 x 10-4(D) + 8.315 x 10-3(A)(B) - 0.022(A)(D) By using experimental data trial #1, the predicted CS was calculated as follows; Sqrt(CS) = +1.05 + 0.071(A) + 8.814 x 10-4(B) + 0.094(C) - 6.955 x 10-4(D) + 8.315 x 10-3(A)(B) - 0.022(A)(D) A = -ve (low) B = -ve (low) C = -ve (low) D = -ve (low) Sqrt(CS) = 1.05 + 0.071(-1) + 8.814 x 10-3(-1) + 0.094(-1) -6.955 x 10-4(-1) + 8.315 x 10-3(-1)(-1) - 0.022(-1)(-1) = 0.7451 mm/min #

From the experimental data trial #1 results, the value of CS is 0.733 mm/min. So the margin error is 2.25%.

Full verification results of the mathematical models for all responses are shown in Table 4.22.

Table 4.22: Margin of error for actual results and predicted values (%) Responses (Predicted)

Exp.

Responses (Actual)

No.

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

Ra ( m) 2.11 1.74 1.90 2.18 2.59 2.45 2.33 2.64 1.94 1.98 2.00 2.00 2.60 2.31 2.72 2.14 1.96 2.07 2.33 2.31

SG (mm) 0.028 0.029 0.029 0.035 0.040 0.040 0.037 0.039 0.032 0.030 0.031 0.029 0.036 0.039 0.041 0.038 0.027 0.030 0.029 0.030

MRR CS (g/min) (mm/min) 0.0042 0.733 0.0062 1.045 0.0050 0.772 0.0082 1.139 0.0093 1.130 0.0125 1.523 0.0084 1.111 0.0128 1.604 0.0055 0.838 0.0060 0.987 0.0054 0.850 0.0062 1.056 0.0089 1.207 0.0112 1.399 0.0098 1.172 0.0114 1.460 0.0077 1.392 0.0084 1.379 0.0082 1.389 0.0084 1.370

Ra ( m) 1.98 1.98 1.98 1.98 2.47 2.47 2.47 2.47 1.98 1.98 1.98 1.98 2.47 2.47 2.47 2.47 2.23 2.23 2.23 2.23

% (error) -6.5 12.4 4.4 -9.8 -4.8 1.1 5.8 -6.8 2.3 0.3 -0.7 -0.9 -5.0 6.8 -9.8 13.7 12.0 7.0 -4.4 -3.7

SG (mm) 0.0304 0.0304 0.0304 0.0304 0.0388 0.0388 0.0388 0.0388 0.0304 0.0304 0.0304 0.0304 0.0388 0.0388 0.0388 0.0388 0.0346 0.0346 0.0346 0.0346

% (error) 8.4 6.2 4.5 -14.7 -3.9 -2.4 4.1 -1.1 -6.7 2.1 -3.2 3.7 6.9 -1.3 -4.5 2.2 22.6 12.0 16.8 14.6

MRR (g/min) 0.0044 0.0076 0.0044 0.0076 0.0091 0.0123 0.0091 0.0123 0.0051 0.0064 0.0051 0.0064 0.0098 0.0111 0.0098 0.0111 0.0082 0.0082 0.0082 0.0082

% (error) 4.0 18.2 -14.3 -8.3 -2.5 -1.8 7.4 -4.3 -8.9 5.5 -6.9 2.4 8.7 -1.4 -0.5 -3.2 6.0 -2.6 -0.2 -2.6

CS (mm/min) 0.754 1.074 0.756 1.146 1.117 1.500 1.119 1.585 0.828 0.984 0.830 1.053 1.206 1.393 1.209 1.475 1.112 1.112 1.112 1.112

% (error) 2.8 2.6 -2.2 0.6 -1.2 -1.5 0.7 -1.2 -1.2 -0.3 -2.4 -0.4 -0.1 -0.4 3.0 1.0 -25.2 -24.0 -24.9 -23.2

82

83

CHAPTER FIVE

DISCUSSIO

5.1

Introduction

The main purpose of this research was to study the effect of WEDM performance on Inconel 718 by using various setting of selected parameter. Performance of the WEDM on Inconel 718 is determined in terms of machining outputs such as surface roughness (Ra), sparking gap (Gap), material removal rate (MRR) and cutting speed (CS). This chapter

elaborates more clearly about the relationship between the performance measures to the main parameters and its influence.

84 5.2 Surface Roughness (Ra)

Based on observation from the ANOVA result for surface roughness (Ra), pulse on (ON) is the only significant factor that influence the Ra. It was revealed when ON was set from 0.65 s up to 0.75 s the Ra was increases about 24.8%. The 3D surface graph for Ra is given in Figure 5.1. Results show that the surfaces profile was in accordance to the model fitted. It was understood that generally Ra increases only by a single factor, ON. Factor of OFF clearly shows that there is no effect on Ra even though the value of OFF was varied from 4 s to 8 s.

Figure 5.1: 3D interaction graph for surface roughness (Ra)

Similar trend of Ra behavior was reported by Mas Ayu [27], the Ra increases when ON increases due to the longer time of machining, leading to the higher possibility of re-sparking and localized sparking to occur. In other words, re-sparking can cause poor surface finish since only the initial phase spark contribute to the material removal rate, while the following spark were poorly distributed along the kerf surface, debris and removed particles.

85 Meanwhile, Ahmet Hascalyk and Ulas Caydas [23] concluded that, the increasing pattern of Ra occurred when intense heat was generated during each electrical discharge. Since the greater the discharge energy conducted into the machining zone, the greater the melted depth of the workpiece that is created. Furthermore, greater discharge energy will produce a larger crater, causing a larger surface roughness value on the workpiece.

5.3

Sparking Gap (Gap)

In term of sparking gap (Gap), the significant factor that influences this response is also a pulse on (ON). It was recorded by ANOVA when the pulse on (ON) is increase its led to increasing of sparking gap (Gap). As for this study, the increment of sparking gap was 27% from low level to the higher level setting of ON. From Figure 5.2, it was understood graphically that the Gap values builds up simultaneously with the pulse on increment of pulse on from 0.65 s to 0.75 s. In this case, pulse off (OFF) also clearly showed there is no effect to the Gap even though the value of OFF varies from 4 s to 8 s.

Figure 5.2: 3D interaction graph for sparking gap (Gap)

86 According to Mas Ayu [27] and Mohd Faisal [31], the wider the ON time, the longer the machining to takes place resulting in a wider spark gap. They reported that servo voltage (SV) also contributed to the effect to the Gap on workpiece materials of tungsten carbide [27] and titanium alloy [31]. As for this study, WEDM Inconel 718 indicated no significant interaction of servo voltage (SV) was recorded on ANOVA.

S.S. Mahapatra and Amar Patnaik [35] also suggested that factor like pulse on (ON) have been found to be significant in effect on sparking gap (Gap). This study revealed similar pattern of Gap behavior when ON is vary from low to the high value. This is due to the increment of power density for the wire to discharge sparks and to elevate the temperature in the gap, hence the higher the power the larger the sparking gap (Gap.

5.4

Material Removal Rate (MRR)

Results obtained from the ANOVA in Table 4.9 clearly show that the most significant factors in affecting material removal rate (MRR) were servo voltage (SV), pulse on (ON) and pulse off (OFF). At the same time, interaction between servo voltage and pulse off (SV*OFF) were also observed to be the significant (Prob>F 0.0076) interaction in this study. Apparently, results obtain from Figure 5.3 indicated that SV and ON contribute to the increment of MRR about 31% and 81% respectively. Meanwhile, the opposite effect of MRR was obtained from the OFF value, whereby MRR was slightly decreased about 2.4% when OFF is varied from 4 s to 8 s. In addition, SV and ON are also interacted to each other which contribute to the increase on MRR. When SV and OFF were set at 60V and 8 s respectively the MRR increase about 17.6%. Meanwhile, MRR increased more rapidly up to 47.8% with an increment of SV when OFF was set at 4 s.

87

Figure 5.3: 3D interaction graph of SV*OFF for material removal rate (MRR)

This result corresponds to the previous researchers finding of Kuang-Yuan Kung and Ko-Ta Chiang [37] whereby they reported that, the electrical spark-erosion process occurs successively and then the removal of melt results in the form of crater on the machined surface. The amount of melt removal determines the level of material removal rate (MRR). The addition of zinc coated in electrode wire provide significantly increase the tensile strength, lowers the melting point and increase the vapor pressure rating resulting in higher MRR. M.S Hewidy et al.[4] proposed, that the increment in the rate of the heat energy hence in the rate of melting and evaporation. Increase in peak current above a certain limit, leads to arcing which decreases the discharge number and the machining efficiency, and subsequently decreases the MRR. Meanwhile, increase in ON time means applying the same heating temperature for a longer time. This will cause an increase in the evaporation rate and number of gas bubbles, which explodes with high ejecting force when the discharge ceases causing removal of bigger volume of the molten metal. Increasing of MRR is continued with the increase of the ejecting force until reaching a situation in which the ejecting force will have no more increase in MRR since the molten metal decreases.

88 5.5 Cutting Speed (CS)

Figure 5.4 presents the 3D effect of cutting speed at various setting of servo voltage (SV) and peak current. The ANOVA results shown in Table 4.10, has shown that the significant parameters for CS were SV, IP and ON. Pulse off (OFF) was only significant during the interaction with SV. However, before ANOVA analysis can be proceed, all the data for CS were obtained from the calculation by dividing the machining distance, d with the machining time, t.

It can be seen that increasing in IP seemly not affected much on CS. From the calculation, CS only increase about 3.45% with increase in IP from 8A up to 12A. Improvement in CS increases dramatically as shown in Figure 5.4 during interaction between IP and SV. Increase in SV from 30V to 60V while IP is maintain at 8A, resulting in increment in CS about 27%. Otherwise, 35% of increment in CS were recorded when SV is varied from 30V up to 60V during maintaining the IP value at 12A. From this finding, the best setting for maximum CS are set at SV = 60V, IP = 12A and ON = 0.75 s.

Figure 5.4: 3D interaction graph of IP*SV for cutting speed (CS)

89 Second interaction that had a significant effect on CS with value of Prob>F below 0.05 was interaction between OFF*SV. When SV is varied between 30V and 60V with a constant OFF of 8 s, the CS experienced a 20.6% increment.

Figure 5.5 indicates graphically a similar pattern was recorded as OFF is set at low value of 4 s with the same setting of SV. It contributes about 43.1% of increment in the CS value. In this case, to achieve maximum value of CS, the SV and OFF should be set at 60V and 4 s respectively.

Figure 5.5: 3D interaction graph of OFF*SV for cutting speed (CS)

Analogously, higher OFF leads to a lower machining time and reduces the CS. This may be due to the fact that during OFF time, the operating impulse was switched off and no current flow at this stage. Too long off time will increase the machining time and reduced CS simultaneously [27]. Based on Figure 5.5, it was obvious that highest CS can dramatically be achieved by setting OFF at low level while SV is set at high level. This setting condition is able to maximize the time for machining and increase the CS. Sufficient setting of OFF time is very important because it can lead to erratic cycling and retraction of the advancing servo, thus slowing down the operation cycle [51].

90

CHAPTER SIX

CO CLUSIO S A D RECOMME DATIO S

6.1

Conclusions

Inconel 718 is a high strength thermal resistant material alloy. It is also a highly strain rate sensitive material which work hardens readily, and contains hard particles making it a very difficult-to-cut material. As for this research, Inconel 718 was machined by using Sodick WEDM linear motor series AQ537L using zinc coated brass wire diameter 0.25mm as the electrode. This research presents an investigation on the effect of machining parameters on WEDM in terms of surface roughness (Ra), sparking gap (Gap), material removal rate (MRR) and cutting speed (CS). The level of importance of the machining parameters on the machining responses was determined by using ANOVA. A total of 20 runs of experiment including centre point were performed in this study which done using Design Expert software version 7.0.0. The following conclusions were drawn based on the performance of machining responses namely; surface roughness (Ra), sparking gap (Gap), material removal rate (MRR) and cutting speed (CS).

91 a. The results of ANOVA and comparisons of experimental data proved that the mathematical models of the value of Ra, Gap, MRR and CS were fairly well fitted with the experimental values with a 95% confidence level. b. The confirmation test show that the errors associated with Ra, Gap, MRR and CS are within the range of 1.54% ~ 12.82%. c. Pulse on (ON) was found to be the most significant factor influencing all responses investigated. Increasing in ON will lead to the low quality of machining responses such as Ra and Gap. Meanwhile, the opposite were observed for MRR and CS whereby increasing of ON will result in better rate of MRR and CS. d. Higher value of thermal conduction and specific heat capacity of Inconel 718 causes the decrease of efficiency of WEDM using zinc coated brass wire as electrode.

6.2

Recommendations

Based on the observation and finding in this study, the future works might attempt to consider the other performance criteria proposed as follows:

a.

The used of different type of wire materials as electrode need to be considered for better understanding for WEDM of Inconel 718.

b.

Surface integrity study can be evaluated in order to understand the effect of the machining parameters on the surface quality and microstructure of the machined surface.

c.

Consideration of others performance criteria, such as surface waviness, form accuracy and surface flatness as additional output parameter fors WEDM of Inconel 718.

92 REFERE CE

1.

Norliana Mohd Abbas, Darius G. Solomon, Md. Fuad Bahari. (2007). A Review on Current Research Trends in Electrical Discharge Machining (EDM). International Journal of Machine Tools & Manufacture. 47: pg. 1214 1228.

2.

J. Proshka, A.G Mamalis, N.M. Vaxevanidis. (1997). The Effect of Electrode Material on Machinability in Wire Electro Discharge Machining. Journal of Material Processing Technology. 69: pg. 233-237.

3.

Lin, Y.C., Yan, B.H., and Chang, Y.S. (2000). Machining Characteristics of Titanium Alloy (Ti-6Al-4V) Using a Combination process of EDM and USM. Journal of Material Processing Technology. 104: pg. 171-177.

4.

M.S. Hewedy, T.A. El-Taweel, M.F. El-Safty. (2005) Modelling the Machining Parameters of Wire Electrical Discharge Machining of Inconel 601 Using RSM. Journal of Material Processing Technology. 169: pg. 328-336.

5.

S. Sharif. (1999). Face Milling of Titanium Alloys Using Coated and Uncoated Carbide Tools. University of Conventry, UK: PhD Thesis.

6.

H. Ho, S.T Newman, S. Rahimifard, R.D. Allen. (2004). State of the Art in Wire Electrical Discharge Machining (WEDM). International Journal of Machine Tool Manufacture 44: pg. 1247-1259.

7.

M. Rahman, W.K.H. Seah, T.T. Teo. (1997). The Machinability of Inconel 718. Journal of Material Processing Technology 63 (1997) 199-204.

93 8. J.C. Rebelo, A. Dias Morao, D. Kremer, J.L. Lebrun. (1998). Influence of EDM Pulse Energy on the Surface Integrity of Martensitic Steel. Journal of Material Processing Technology 84: pg. 9096.

9.

Rival. (2005). Electrical Discharge Machining of Titanium Alloy Using Copper Tungsten Electrode With SiC Powder Suspension Dielectric Fluid. Universiti Teknologi Malaysia: Master Thesis.

10.

I. Puertas, C.J. Luis. (2003). Modeling the Manufacturing Parameters in Electrical Discharge Machining of Siliconized Silicon Carbide. Proc. Inst. Mech., J. Eng. Manuf. 217 (Part B): pg. 791803.

11.

S. Sarkar, S. Mitra, B. Bhattacharyya. (2005). Parametric Analysis and Optimization of Wire Electrical Discharge Machining of Titanium Aluminide Alloy. Journal of Materials Processing Technology 159: pg. 286294.

12.

Sommer, C. (2000). Non-Traditional Machining Handbook. First edition. Advanced Publishing Houston.

13.

Pandey, P.C., and Shah, H.S. (1980). Modern Machining Processes. New Delhi. Tata Mcgraw-Hill.

14.

K.P. Rajurkar. (1994). Handbook of Design, Manufacturing and Automation, Chapter 13: Nontraditional Manufacturing Processes. Wiley, USA.

15.

George F. Schrader, Ahmad K. Elshennawy, Lawrence E. (2000). Manufacturing Processes & Materials. Forth Edition. SME, ISBN: 0872635171.

16.

Helmi A. Youssef, Hassan El-Hofy. (2008). Machining Technology Machine Tools and Operation, Chapter 11: Nontraditional Machine Tools and operation. CRR Press. ISBN: 978-1-4200-4339-6.

94 17. Jack M. Walker. (1996). Hanbook of Manufacturing Engineering, Chapter 16: Nontraditional Machining Methods. Marcel Decker, Inc., ISBN: 0-82478962-8. 18. Mchado A.R. and J. Wallbank. (1990). Machining of Titanium and its Alloys A Review. Proc. Instn. Mech. Engrs. Vol. 204: pg. 53-60. 19. Guitrau, E.B. (1991). Wire EDM An Overview of Technology and Trends. SME Technical Paper, MR 91-519. 20. William, R.E., Rajurkar, K.P. (1991). Study of Wire Electrical Discharge Machining Surface Characteristic. Journal of Materials Processing Technology, Vol. 28: pg. 486-493. 21. Jerzy Kozak, Kamlakar P. Rajukar, Niraj Chandarana. (2004). Machining Low Electrical Conductive Materials by Wire electrical Discharge Machining (WEDM). Journal of Materials Processing Technology 149: pg. 266271. 22. D. Scot, S. Boyina, K.P. Rajurkar. (1991). Anallysis and Optimization of Parameters Combination in Wire Electrical Discharge Machining. Int. J. Prod. Res. 29 (11): pg. 2189-2207. 23. Ahmet Hascalyk, Ulas Caydas. (2004). Experimental Study of Wire Electrical Discharge Machining of AISI D5 Tool Steel. Journal of Material Processing Technology 148: pg. 362-367. 24. Y.S. Liao, J.T. Huang, H.C. Su. (1997). A Study on the Machining Parameters Optimization of Wire Electrical Discharge Machining. Journal of Material Processing Technology 71: pg. 487-493. 25. Wang, C.C., and. Yan, B.H. (2000). Blind Hole Drilling of Al2O3/6061 Al Composite Using Rotary Electro Discharge Machining. Journal of Material Processing Technology 102: pg. 90-102.

95

26.

Van Tri, N. (2002). Electrical Discharge Machining of Aluminum Alloy Using Classical Design of Experiment Approach. Universiti Teknologi Malaysia: Master Thesis.

27.

Mas Ayu. (2006). Investigation of WEDM Machining Tungsten Carbide. Universiti Teknologi Malaysia: Master Thesis.

28.

Nihat Tosun, Can Cogun, Gul Tosun. (2004). A Study on Kerf and Material Removal Rate in Wire Electrical Discharge Machining Based on Taguchi Method. Journal of Material Processing Technology 152: pg. 316-322.

29.

Erden, A. (1982). Role of Dielectric Flushing on Electric Discharge Machining Performance. Machine Tool Design and Research Conference (1982) September 14-15. Manchester. 23: pg. 283-289.

30.

Kinoshita, N., Fukui, M., Shichida, H., Gamo, G., sata, T. (1976). Study on EDM with Wire Electrode; Gap Phenomena. Ann. CIRP 25 (1976) (1): pg. 141-145.

31.

Faisal Othman. (2008). EDM Wirecut of Titanium Alloy (Ti-6A14V) Using Brass Coated Wire Electrode. Universiti Teknologi Malaysia: Master Thesis.

32.

O. A. Abu Zeid. (1997). On the Effect of Electro Discharge Machining Parameter on the Fatigue Life of AISI D6 Tool Steel. Journal of Material Processing Technology 68: pg. 27-32.

33.

H.T. Lee, T.Y. Tai. (2003). Relationship between EDM Parameters and Surface Crack Formation. Journal of Material Processing Technology 142: pg. 676-683.

34.

Q. H. Zhang, J.H. Zhang, S.F. Ren, Z.W. Niu, X. Ai. (2005). A Theoritical Model of Surface Roughness in Ultrasonic Vibration Assisted Electrical

96 Discharge Machining in Gas. International Journal of Manufacturing Technology and Management 7: 381-390. 35. S.S Mahapatra, Amar Panaik. (2007). Optimization of Wire Electrical Discharge Machining (WEDM) Using Taguchi Method International Journal of Manufacturing Technology 34: pg. 911-925. 36. S. Sarkar, M. Sekh, S. Mitra, B. Bhattacharyya. (2008). Modelling and Optimization of Wire Electrical Discharge Machining of -TiAl in Trim Cutting Operation. Journal of Material Processing Technology 205: pg. 376-387. 37. Kuang-Yuan Kung, Ko-Ta Chiang. (2008). Modeling and Analysis of Machinability Evaluation in the Wire Electrical Discharge Machining (WEDM) Process of Aluminum Oxide-Based Ceramic. Material and Manufacturing Processes 23: pg. 241-250. 38. R. Ramakrishnan, L. Karunamoorthy. (2008). Modelling and Multi-Responses Optimization of Inconel 718 on Machining of CNC WEDM Processes. Journal of Material Processing Technology 207: pg. 342-349. 39. Z.N. Guo, T.C. Lee, T.M. Yue, W.S. Lau. (1997). A Study of Ultrasonicaided Wire Electrical Discharge Machining. Journal of Materials Processing Technology 63: pg. 823828. 40. M. Kunieda, S. Furuoya, N. Taniguchi. (1991). Improvement of EDM Efficiency by Supplying Oxygen Gas into gap. CIRP Annals - Manufacturing Technology 40: pg. 215218. 41. M.L. Jeswani. (1981). Effect of the Addition of Graphite Powder to Kerosene Used as the Dielectric Fluid in Electrical Discharge Machining. Wear 70: pg. 133139.

97 42. Mikell P. Groover. (2007). Fundamentals of Modern Manufacturing Materials, Process and Systems, Chapter 26. John Wiley & Sons, Inc. ISBN: 13978-0-74485-6. 43. Intech EDM. (1995). EDM Wire: A Reference to Understanding, Selecting and Using Wire on Wire-Cut EDM Machine. 2nd Edition USA. 44. Armarego, E.J.A., and Brown, R.H. (1969). The Machining of Metals. Prentice Hall New Jersey. 45. Mustafa Ihan Gokler, Alp Mihat Ozanozgu. (1997). Experimental Investigation of Effects of Cutting Parameters on Surface Roughness in the WEDM Process. 46. F.C. Campbell. (2006). Manufacturing Technology for Aerospace Structural Materials, Charter 6: Super Alloys. Elsevier Ltd. ISBN-13: 978-1-85617495-4. 47. I.A. Choudhury, M.A. El-Baradie. (1999). Machinability Assessment of Inconel 718 by Factorial Design of Experiment Coupled with Response Surface Methodology. Journal of Material Processing Technology 95: pg. 30-39. 48. Graham, D. (2002). Turning Difficult-to-Machined Alloy. Modern Machine Shop. 49. Lochner, R.H., and Matar, J.E. (1990). Designing for Quality; an Introduction to the Best of Taguchi and Western Methods of Statistical Experiment Design. Chapman and Hall, New York. 50. Montogory. (2001). Practical Experiment Design for Engineers and Scientist. 3rd Edition, John Waley & Sons, Inc. New York.

98 51. Lee, S.H. and Li, X.P. (2001). Study of the Effect of Machining Parameters on the Machining Characteristics in Electrical Discharge Machining of Tungsten Carbide. Journal of Materials Processing Technology. 115: pg. 344-358.

99

APPE DIX A MASTER PROJECT PLA I G

100

Appendix A1: Schedule for Master project part I (Semester 1 2009/2010)

100

Appendix A2: Schedule for Master project part II (Semester 2 2009/2010)

101

102

APPE DIX B SUMMARY OF FI DI G RELATED TO EDM PERFORMA CE

103

Appendix B1: Summary of finding related to EDM performance

Researcher

Issue and Methodology Used Development of mathematical model for correlating the interrelationship of various WEDM machining parameter (peak current, duty factor, wire tension, and water pressure) at MRR, spark gap & surface finish. Material: Inconel 601 (6mm Thickness) Electrode: Brass Wire CuZn377, .25mm Method: RSM Optimization of WEDM process parameter (WEDM parameters: discharge current, pulse duration, pulse frequency, wire speed, wire tension and dielectric flow) Material: AISI D2 tool steel (10mm thickness) Electrode: Zink-coated cooper wire (Stratified, copper .25mm) Method: Taguchi Method Modeling and optimization of WEDM in trim cutting operation (WEDM parameters:cutting speed). Material: -TiAl (15mm thickness) Electrode: Brass wire 0.25mm Method: RSM

Result/Conclusion Result show surface finish greatly influence by peak current, duty factor and wire tension. Surface finish increase by increase in peak current whereby decrease with the increase in duty factor and wire tension.

M.S. Hewidy et al. [4].

S.S Mahapatra & Amar Panaik [35]

Determine the discharge current, pulse duration and dielectric flow rate as significant role in rough cutting to minimize surface finish and increase the MRR.

Determine the surface finish decrease as the cutting speed increase.

S. Sarkar et al. [36]

104 Modeling and analysis of machinibility evaluation in the WEDM process of aluminum oxide-based ceramic (WEDM parameters: peak current, pulse-on time, duty factor and wire speed). Material: Aluminium oxide Al2O3 (10mm thickness) Electrode: Cylindrical electrolytic copper, 0.20mm. Method: RSM Modeling and multi-responses optimization of Inconel 718 on machining of CNC WEDM process (WEDM parameters: pulse on-time, pulse off-time, wire feed speed & ignition current) Material: Inconel 718 (14mm) Electrode: Brass wire, 0.25mm Method: Taguchi Method. Study on kerf and MRR in WEDM machining proceses (WEDM parameters: pulse duration, open circuit voltage, wire speed & dielectric pressure) Material: AISI 4140 steel (10 mm thickness) Electrode: Brass wire CuZn37, 0.25mm Method: Taguchi Methods Conclude the values of MRR and surface finish increase with the increase in pulse on-time and duty factor up to certain limits then decrease with the further increase in the pulse on-time and duty factor.

Kuang-Yuan Kung & Ko-Ta Chiang

R. Ramakrishnan & L. Karunamoorthy [38]

Results shows by an increase of pulse time and ignition current effect on MRR was improved. But at higher rates of pulse on-time and ignition current, the surface finish of the Inconel 718 was affected.

Nihat Tosun et al. [28]

Revealed the highly effective parameters on both kerf and MRR were found as open circuit voltage and pulse duration. Wire speed & dielectric pressure were less effective factors.

105

APPE DIX C EXPERIME TAL RESULTS OF SPARKI G GAP (TOP A D BOTTOM SURFACE)

106 Appendix C1: Experimental results of sparking gap (top surface) Kerf Width Average (mm) 0.313 0.312 0.314 0.328 0.335 0.325 0.327 0.328 0.322 0.317 0.320 0.317 0.321 0.325 0.335 0.345 0.324 0.322 0.314 0.314 Sparking Gap Average (mm) 0.031 0.031 0.032 0.039 0.043 0.038 0.039 0.039 0.036 0.034 0.035 0.034 0.036 0.037 0.042 0.047 0.037 0.036 0.032 0.032

Exp. No. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

Kerf Width (mm) 1 0.314 0.307 0.319 0.328 0.334 0.320 0.322 0.328 0.321 0.318 0.327 0.329 0.326 0.336 0.343 0.371 0.319 0.320 0.308 0.318 2 0.309 0.303 0.315 0.328 0.338 0.331 0.340 0.323 0.317 0.314 0.316 0.326 0.315 0.315 0.332 0.342 0.316 0.317 0.308 0.303 3 0.315 0.326 0.307 0.327 0.334 0.325 0.320 0.334 0.329 0.319 0.317 0.297 0.323 0.323 0.329 0.321 0.338 0.330 0.325 0.322

Sparking Gap (mm) 1 0.032 0.029 0.035 0.039 0.042 0.035 0.036 0.039 0.036 0.034 0.039 0.040 0.038 0.043 0.047 0.061 0.035 0.035 0.029 0.034 2 0.030 0.027 0.033 0.039 0.044 0.041 0.045 0.037 0.034 0.032 0.033 0.038 0.033 0.033 0.041 0.046 0.033 0.034 0.029 0.027 3 0.033 0.038 0.029 0.039 0.042 0.038 0.035 0.042 0.040 0.035 0.034 0.024 0.037 0.037 0.040 0.036 0.044 0.040 0.038 0.036

107 Appendix C2: Experimental results of sparking gap (bottom surface) Kerf Width Average (mm) 0.299 0.302 0.302 0.312 0.326 0.333 0.321 0.328 0.307 0.302 0.305 0.300 0.323 0.332 0.327 0.307 0.283 0.299 0.301 0.304 Sparking Gap Average (mm) 0.024 0.026 0.026 0.031 0.038 0.042 0.036 0.039 0.029 0.026 0.028 0.025 0.037 0.041 0.039 0.029 0.016 0.025 0.026 0.027

Exp. No. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

Kerf Width (mm) 1 0.297 0.315 0.291 0.304 0.319 0.325 0.316 0.318 0.305 0.297 0.297 0.295 0.319 0.327 0.331 0.286 0.271 0.300 0.296 0.287 2 0.295 0.290 0.310 0.308 0.330 0.338 0.321 0.338 0.319 0.301 0.305 0.296 0.329 0.334 0.323 0.313 0.291 0.295 0.308 0.316 3 0.304 0.301 0.306 0.323 0.328 0.337 0.327 0.329 0.298 0.308 0.314 0.308 0.321 0.336 0.328 0.322 0.286 0.303 0.300 0.308

Sparking Gap (mm) 1 0.024 0.033 0.021 0.027 0.035 0.038 0.033 0.034 0.028 0.024 0.024 0.023 0.035 0.039 0.041 0.018 0.011 0.025 0.023 0.019 2 0.023 0.020 0.030 0.029 0.040 0.044 0.036 0.044 0.035 0.026 0.028 0.023 0.040 0.042 0.037 0.032 0.021 0.023 0.029 0.033 3 0.027 0.026 0.028 0.037 0.039 0.044 0.039 0.040 0.024 0.029 0.032 0.029 0.036 0.043 0.039 0.036 0.018 0.027 0.025 0.029

108

APPE DIX D EXPERIME TAL RESULTS OF SPARKI G GAP FOR CO FIRMATIO TEST (TOP A D BOTTOM SURFACE)

109 Appendix D1: Confirmation experimental results of sparking gap (top surface) Kerf Width Average (mm) 0.313 0.308 0.313 Sparking Gap Average (mm) 0.032 0.029 0.031

Exp. No. 1. 2. 3.

Kerf Width (mm) 1 0.313 0.313 0.310 2 0.318 0.312 0.313 3 0.309 0.300 0.315

Sparking Gap (mm) 1 0.032 0.032 0.030 2 0.034 0.031 0.032 3 0.030 0.025 0.033

Appendix D2: Confirmation experimental results of sparking gap (bottom surface) Kerf Width Average (mm) 0.313 0.308 0.313 Sparking Gap Average (mm) 0.032 0.029 0.031

Exp. No. 1. 2. 3.

Kerf Width (mm) 1 0.313 0.313 0.310 2 0.318 0.312 0.313 3 0.309 0.300 0.315

Sparking Gap (mm) 1 0.032 0.032 0.030 2 0.034 0.031 0.032 3 0.030 0.025 0.033